Forecasting Water Supply with B120 Statistical Methodswith B120 Statistical MethodsExample:Cosumnes River

California Cooperative Meeting 11/8/12California Cooperative Meeting  11/8/12By John j. King PE

AA

Water Year 20128 Station Index8 Station Index

Cosumnes Michigan BarB120A‐J % Error FNFCosumnes, Michigan Bar A‐J   % Error FNF

April 1 Forecast 50 taf ‐52% 105taf

Cosumnes, Michigan BarB120A‐J  % Error FNF

April 1 Forecast 50taf ‐52% 105taf

May 1 Forecast 100taf ‐5% 105taf

Cosumnes, Michigan BarB120A‐J  % Error FNF

April 1 Forecast 50taf ‐52% 105tafp

May 1 Forecast 100taf ‐5% 105taf

End of Season Verification 59taf ‐44% 105taf

ROA‐J=.776S4/1 +.094POM+.0073POM2+.00362PAJ2

2012 S4/1=8.8 in POM=23.7 in PAJ=6.6 in

1961 1964 1972 1990 2001

October‐March Precipation Total

30 0

35.0

p

29.4   Average

22.9

27.9 27.824.1 24.2 23.725.0

30.0

20.0

10.0

15.0in

5.0

0.01961 1964 1972 1990 2001 2012

Year

April‐June Precipation Total

7.47.1

6 67.0

8.0

April‐June Precipation Total

5.2 5.25 7

6.6

6.0

7.0

5.9   Average

5.7

4.0

5.0

i

2.0

3.0in

1.0

2.0

0.01961 1964 1972 1990 2001 2012

Year

A il 1 SWE

18.0

20.0April 1 SWE

19.9 Average

14.0

16.0

11.9

8 6 8 6 8.810.0

12.0

E  in.

6.2

8.6

7.6

8.6

6.0

8.0SWE

2.0

4.0

0.01961 1964 1972 1990 2001 2012

Year

% A‐J Forecast Error (End of Season)

40%

50%

20%

30%

10%

20%

‐10%

0%1961 1964 1972 1990 2001 2012

‐20%

‐40%

‐30%

October‐June Precipation Distribution12.0

October June Precipation Distribution

8.0

10.0

15.4 inches fell in March and April 50% of October‐June Total

6.0

8.0

in

2012Average

4.0

2.0

0.0Oct Nov Dec Jan Feb Mar Apr May Jun

Month

P i it ti

Solarradiation

Physical Properties Modeled with B120 Statistical Model

Plant canopyi t ti

Precipitation

Air temperatureSublimationEvaporation

interception

Surface runoff

Snowpack

Transpiration

Evaporationand

TranspirationEvaporation

RainRain

Snowmelt

Throughfall

Streamflowto stream

Impervious-Zone ReservoirSoil-Zone Reservoir

p Snowmelt

Recharge zoneLower zone

Streamflow

SubsurfaceReservoir

Subsurface recharge

Ground-waterrecharge

Interflow orsubsurface

flow to stream

Ground-water recharge

Ground-WaterReservoir Ground-water flow to stream

flow to stream

Ground-watersink

Benefits of PRMS Modeling: Feather River SubsurfaceFeather River Subsurface 

Representation

Khalida FazelPRMS Workgroup

Feather River Basin

Feather PRMS Flow Processes at a ll lHillslope Cross‐Section

Source: USGS GSFlow Manual

Feather PRMS Moisture Accounting ConceptualizationConceptualization

Feather PRMS Precipitation Partitioning, OR HRU 19 WY 2012

4OR HRU 19 WY 2012

3

2Dep

th (in)

Precip

Precip

Snowfall

1

00Oct‐11 Nov‐11 Dec‐11 Jan‐12 Feb‐12 Mar‐12 Apr‐12

OR HRU 19 PRMS Moisture Accounting WY 20120.001Oct2011 01Nov2011 01Dec2011 01Jan2012 01Feb2012 01Mar2012 01Apr2012 01May2012 01Jun2012 01Jul2012 01Aug2012

2.5

Precip(in

)

5.0

0 0

2.5

5.0

nowmelt 

(in)

Field capacity = 8.74 in

0.0S

5.0

7.5

10.0

oil Zon

e eservoir 

orage (in

)

0.0

2.5S R Sto

5 0

7.5

face 

oir 

ge 

0.0

2.5

5.0

Subsurf

Reservo

Storag

(in)

3.0

ter 

r n)

0.0

1.5

Groun

dwat

Reservoir

Storage (in

OR Subsurface Reservoir Outflow WY 2012ace Ru

noff 

(cfs)

Surfa

surface 

w (cfs)

ter 

)Subs

Flow

w 

Groun

dwat

Flow

 (cfs

Stream

flow

Flow

 (cfs)

OR Streamflow Components WY 2012OR Streamflow Components WY 2012

Streamflow

Baseflow

Forecasting with PRMS: Antecedent ConditionsAntecedent Conditions

Source: EM 1110‐2‐1406

ConclusionsConclusions

• Knowledge of storage in each subsurface reservoir o edge o sto age eac subsu ace ese oand the ability to quantify input as either rain or snowmelt can better inform forecast

• PRMS provides a representation of streamflow and its contributing components

• Extreme/unusual events can be better represented in a distributed, conceptual model (PRMS), compared to statistical approaches (Bulletin 120)compared to statistical approaches (Bulletin 120)

Lower N th F kNorth Fork

03000

MMS run for new Bucks Lake subbasin, WY 1996-1999

32500

62000

atio

n (in

ches

)

cks

Lake

(cfs

)

91500

Ave

rage

Pre

cipi

ta

ek a

t Low

er B

uc

Initial Simulation

Observed (NF904)

12

15500

1000

Subb

asin

Buc

ks C

re

15

180

500

0150n-ppt (in.) melt (in.) pweqv (in.) Mill Creek Flat, 5900' Bucks Lake SWE, 5750' Letterbox, 5600'

1

2

3120

130

140

4

5

690

100

110

t wat

er (i

nche

s)

ent (

inch

es)

7

8

960

70

80

cipi

tatio

n &

mel

t

ow w

ater

equ

ival

10

11

1230

40

50

Net

pre

Sno

13

14

150

10

20

995

996

997

998

999

000

001

002

003

004

005

006

007

008

009

010

011

7/1/

19

7/1/

19

7/1/

19

7/1/

19

7/1/

19

7/1/

20

7/1/

20

7/1/

20

7/1/

20

7/1/

20

7/1/

20

7/1/

20

7/1/

20

7/1/

20

7/1/

20

7/1/

20

7/1/

20

0150n-ppt (in.) melt (in.) pweqv (in.) Adjust lapse rate Mill Creek Flat, 5900' Bucks Lake SWE, 5750' Letterbox, 5600'

1

2

3120

130

140

4

5

690

100

110

t wat

er (i

nche

s)

ent (

inch

es)

7

8

960

70

80

cipi

tatio

n &

mel

t

ow w

ater

equ

ival

10

11

1230

40

50

Net

pre

Sno

13

14

150

10

20

995

996

997

998

999

000

001

002

003

004

005

006

007

008

009

010

011

7/1/

19

7/1/

19

7/1/

19

7/1/

19

7/1/

19

7/1/

20

7/1/

20

7/1/

20

7/1/

20

7/1/

20

7/1/

20

7/1/

20

7/1/

20

7/1/

20

7/1/

20

7/1/

20

7/1/

20

03000

MMS run for new Bucks Lake subbasin, WY 1996-1999

32500

62000

atio

n (in

ches

)

cks

Lake

(cfs

)

91500

Ave

rage

Pre

cipi

ta

ek a

t Low

er B

uc

Initial SimulationObserved (NF904)Adjust lapse rate

12

15500

1000

Subb

asin

Buc

ks C

re

Adjust lapse rate

15

180

500

PRMS Input Stations

Canyon Dam

Q iB k C k

Caribou PowerhouseDeSabla

QuincyBucks Creek Powerhouse

Brush Creek 

Strawberry Valley

PRMSPrecipitation Runoff Modeling System

http://wwwbrr.cr.usgs.gov/projects/SW_MoWS/PRMS.html

PRISMPRISMParameter‐elevation Regressions on 

Independent Slopes Modelhttp://www.prism.oregonstate.edu/

Short-term Weather Forecast - March 12 2012

9000

10000

Short-term Weather Forecast - March 12, 2012Observed Simulated Forecast

7000

8000

Ric

h B

ar (c

fs)

5000

6000

her R

iver

nea

r R

3000

4000

anch

of N

F Fe

ath

Observed vs

Pre-event forecast

1000

2000

East

Bra

07/1/2010 10/1/2010 1/1/2011 4/1/2011 7/1/2011 10/1/2011 1/1/2012 4/1/2012 7/1/2012

ESP Forecast as of April 1, 20121970 1971 1972 1973 1974 1975 1976 19771978 1979 1980 1981 1982 1983 1984 19851986 1987 1988 1989 1990 1991 1992 19931994 1995 1996 1997 1998 1999 2000 2001

9000

10000

fs)

1994 1995 1996 1997 1998 1999 2000 20012002 2003 2004 2005 2006 2007 2008 20092010 2011 25% 75% Med Ave Observed

7000

8000

ear R

ich

Bar

(cf

5000

6000

Feat

her R

iver

ne

3000

4000

t Bra

nch

of N

F F

Observed vs

Post-event simulation

1000

2000East

07/1/2010 10/1/2010 1/1/2011 4/1/2011 7/1/2011 10/1/2011 1/1/2012 4/1/2012 7/1/2012

East Branch

Confluence

7 22

23

8

QuincyQuincy

PRMS Feather River PRMS Model Improvement Status

C lif i C i M iCalifornia Cooperative MeetingNovember 8, 2012

Short‐term Forecasting Improvements with Feather PRMS 

ModelModel

CNRFC Gridded QPF Datal bl d hAvailable :3 days, 6 hour segment

QPF 4.7km grid, CDF file format

Geo‐processing Gridded QPF for PRMS HRUsfor PRMS HRUsProcess the CDF file to a ESRI shape file

Add the four 6 hour QPF grids for each day

p

Q g y

Sample the gridded QPF for the centroid of the HRU

Append data to input fileAppend data to input file(Climate Q files)