RasterRaster--based based Streamflow Streamflow

AnalysisAnalysis

Richard B. Koehler, Ph.D.Richard B. Koehler, Ph.D.National Weather Service/NOAANational Weather Service/NOAA

Boulder, COBoulder, CO

BackgroundBackground

Streamflow patterns Streamflow patterns Occur on different timescalesOccur on different timescalesInclude flow volume and timing Include flow volume and timing Show cumulative effect of disturbancesShow cumulative effect of disturbances

Multiple existing methods (170+ indices)Multiple existing methods (170+ indices)Many correlated or redundantMany correlated or redundantAdequate for volume (composition)Adequate for volume (composition)Weak for timing (configuration)Weak for timing (configuration)

Large daily datasets existLarge daily datasets exist

Study sitesStudy sites

Snake River at Irwin, IDat Heise, ID

San Miguel Riverat Placerville, CO

(CONTROL)

Colorado Riverat Lees Ferry, AZ

Palisades Res, WY

Year 1984

Day 329 330 331

...1985

329 330 331

...1986

329 330 331

1 yr 1 yr

Year

1986

1985

1984

Day 329 330 331

Short-term timescale "i" (daily)

Long-term timescale

"j"(yearly)

Raster gridded time seriesRaster gridded time series

Colorado River exampleColorado River exampleColorado River at Lees Ferry, AZ linear hydrographColorado River at Lees Ferry, AZ linear hydrograph

10

100

1000

10000

Oct-1921

Oct-1931

Oct-1941

Oct-1951

Oct-1961

Oct-1971

Oct-1981

Oct-1991

Oct-2001

Date

Flow

(cm

s)

1

2

3

45

Colorado River exampleColorado River exampleColorado River at Lees Ferry, AZ raster hydrographColorado River at Lees Ferry, AZ raster hydrograph

30 60 90 120 150 180 210 240 270 300 330 360

Day of Water Year

1930

1940

1950

1960

1970

1980

1990

2000

Wat

erY

ear

100

1000

Flow (cms)

2.00

3.00

11

22

3

4

5

Autocorrelation Autocorrelation Linear vs. gridLinear vs. grid--based lag schemebased lag scheme

∆t = 1 day lag

Comparison cells

∆t = 1 day lag ∆t = 1 year lag ∆t = 1 year and 1 day lag

CorrelogramsCorrelograms

San Pedro River at San Pedro River at Charleston, AZCharleston, AZ

1936 1936 -- 20012001

-0.2-0.10.00.10.20.30.40.50.60.70.80.91.0

0 30 60 90 120 150 180 210 240 270 300 330 360Lag, k (days)

Cor

rela

tion,

r(k)

DailyDaily

YearlyYearly

0.00.10.20.30.40.50.60.70.80.91.0

0 1 2 3 4 5 6 7 8 9 10Lag, k (years)

Cor

rela

tion,

r(k)

CorrelogramsCorrelogramsr(ki,kj) =

var (unshifted cells)covar (shifted cells)

2

1

Year

ly la

g (j

)

-1

-2

-2 -1 0 1 2

Daily lag (i)

0

1r(k)

0

1

Daily Lags

0

1r(

k)

01

Yea

rly L

ags

0

Artificial flow examplesArtificial flow examplesRandom yearly flowRandom daily flow

30 60 90 120 150 180 210 240 270 300 330 360

"Day"

1900

1920

1940

1960

1980

"Yea

r"

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

"Flow"

30 60 90 120 150 180 210 240 270 300 330 360

"Day"

1900

1920

1940

1960

1980

"Yea

r"

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

"Flow"

-150 -100 -50 0 50 100 150

Lag in "days"

-40

-20

0

20

40

Lag

in "y

ears

"

-0.25

0.00

0.25

0.50

0.75

1.00Correlation

-150 -100 -50 0 50 100 150

Lag in "days"

-40

-20

0

20

40

Lag

in "y

ears

"

-0.25

0.00

0.25

0.50

0.75

1.00Correlation

Artificial flow examples, part 2Artificial flow examples, part 2Random fluctuating daily flowExactly identical increasing yearly flow

50 100 150 200 250 300 350

"Day"

1900

1920

1940

1960

1980

"Yea

r"

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0"Flow""Flow"

30 60 90 120 150 180 210 240 270 300 330 360

"Days"

1900

1920

1940

1960

1980

"Yea

rs"

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

"Flow"

30 60 90 120 150 180 210 240 270 300 330 360

"Days"

1900

1920

1940

1960

1980

"Yea

rs"

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

-150 -100 -50 0 50 100 150

Lag in "days"

-40

-20

0

20

40

Lag

in "y

ears

"

Correlation

-0.25

0.00

0.25

0.50

0.75

1.00

-150 -100 -50 0 50 100 150

Lag in "days"

-40

-20

0

20

40La

g in

"ye

ars"

-0.25

0.00

0.25

0.50

0.75

1.00

Correlation

30 60 90 120 150 180 210 240 270 300 330 360

Day of Water Year

1950

1960

1970

1980

1990

2000

Wat

er Y

ear

0. 0

1. 0

1

10

Flow (cms)

ResultsResultsSan Miguel River San Miguel River at Placerville, COat Placerville, CO

Control site for Control site for Upper Colorado Upper Colorado River Basin

Raster hydrographRaster hydrographRiver Basin

Grid correlogramGrid correlogram

-150 -100 -50 0 50 100 150

Lag in days

-30

-20

-10

0

10

20

30La

g in

yea

rs

-0 .25

0.00

0.25

0.50

0.75

1.00

Correlation

1000

10000

100000

30 60 90 120 150 180 210 240 270 300 330 360

Day of Water Year

196019651970197519801985199019952000

Wat

er Y

ear

3.00

3.50

4.00

4.50

Flow (cms)

ResultsResults

Palisades Palisades ReservoirReservoir

End of month End of month storage converted storage converted to streamflow

Raster hydrographRaster hydrographto streamflow

Grid correlogramGrid correlogram

-150 -100 -50 0 50 100 150

Lag in days

-20

-10

0

10

20La

g in

yea

rs

-0 .25

0.00

0.25

0.50

0.75

1.00

Correlation

10

100

1000

30 60 90 120 150 180 210 240 270 300 330 360Day of Water Year

1920

1930

1940

1950

1960

1970

1980

1990

2000

Wat

er Y

ear

Flow (cms)

Raster hydrographRaster hydrograph

ResultsResultsSnake River at Snake River at Heise, IDHeise, ID

Downstream from Downstream from Palisades ReservoirPalisades Reservoir

-150 -100 -50 0 50 100 150

Lag in days

-20

-10

0

10

20

Lag

in y

ears

-0.25

0.00

0.25

0.50

0.75

1.00

Correlation

(1960 through 2000)(1960 through 2000)

-150 -100 -50 0 50 100 150

Lag in Days

-20

-10

0

10

20

Lag

in y

ears

Correlation

-0.25

0.00

0.25

0.50

0.75

1.00

(1911 through 1951)(1911 through 1951) Grid correlogramsGrid correlograms

10

100

1000

30 60 90 120 150 180 210 240 270 300 330 360

Day of Water Year

1960

1970

1980

1990

2000

Wat

er Y

ear

1

2

3

Flow (cms)

-150 -100 -50 0 50 100 150

Lag in days

-20

-10

0

10

20

Lag

in y

ears

-0.25

0.00

0.25

0.50

0.75

1.00

Correlation

Raster hydrographRaster hydrographGrid correlogramGrid correlogram

ResultsResults

Snake River at Snake River at Irwin, IDIrwin, ID

1956 1956 -- 20022002

Observed Observed hydrograph and hydrograph and gird correlogramgird correlogram

30 60 90 120 150 180 210 240 270 300 330 360

Day of Water Year

196019651970197519801985199019952000

Wat

er Y

ear

3.00

4.00 10000

1000

Flow (cms)

ResultsResults

Snake River at Snake River at Irwin, IDIrwin, ID

1956 1956 -- 20022002

Adjusted Adjusted hydrograph and hydrograph and gird correlogram

Raster hydrographRaster hydrograph

Grid correlogramGrid correlogramgird correlogram

-150 -100 -50 0 50 100 150

Lag in days

-20

-10

0

10

20

Lag

in y

ears

Correlation

-0.25

0.00

0.25

0.50

0.75

1.00

30 60 90 120 150 180 210 240 270 300 330 360

Day of Water Year

1930

1940

1950

1960

1970

1980

1990

2000

Wat

erY

ear

100

1000

Flow (cms)

2.00

3.00

30 60 90 120 150 180 210 240 270 300 330 360

Day of Water Year

1930

1940

1950

1960

1970

1980

1990

2000

Wat

erY

ear

100

1000

Flow (cms)

2.00

3.00ResultsResultsColorado River at Colorado River at Lees Ferry, AZLees Ferry, AZ

Downstream from Downstream from Glen Canyon Dam Raster hydrographRaster hydrographGlen Canyon Dam

(1930 (1930 -- 1960)1960) (1970 (1970 -- 2000)2000)Grid correlogramsGrid correlograms

-150 -100 -50 0 50 100 150

Lag in days

-10

0

10La

g in

yea

rs

Correlation

-0.250.00

0.250.500.751.00

-150 -100 -50 0 50 100 150

Lag in days

-10

0

10

Lag

in y

ears

-0.250.000.250.500.751.00

Correlation

Climate applicationsClimate applicationsTemporal and spatial analysisTemporal and spatial analysis

WY 3

CO 2

UT 6

UT 7

NM 1

NM 4

AZ 2

AZ 3AZ 4

1895

2002

Climate applicationsClimate applications

6 12

NM 4

1900

1910

1920

1930

1940

1950

1960

1970

1980

1990

2000

6 12

NM 1

1900

1910

1920

1930

1940

1950

1960

1970

1980

1990

2000

6 12

AZ 4

1900

1910

1920

1930

1940

1950

1960

1970

1980

1990

2000

6 12

AZ 3

1900

1910

1920

1930

1940

1950

1960

1970

1980

1990

2000

6 12

Month of Cal Year

AZ 2

1900

1910

1920

1930

1940

1950

1960

1970

1980

1990

2000

6 12

UT 6

1900

1910

1920

1930

1940

1950

1960

1970

1980

1990

2000

6 12

UT 7

1900

1910

1920

1930

1940

1950

1960

1970

1980

1990

2000

6 12

CO 2

1900

2000

6 12

WY 3

1900

1910

1920

1930

1940

1950

1960

1970

1980

1990

2000

-8-6-4-202468

PHDI

Wet

Dry

972 years of monthly values

Cal

enda

r Yea

r

Other applicationsOther applications

6 12

UT 6

1900

1910

1920

1930

1940

1950

1960

1970

1980

1990

2000

6 12

UT 7

1900

1910

1920

1930

1940

1950

1960

1970

1980

1990

2000

6 12Month of Cal Year

CO 2

1900

2000

6 12

WY 3

1900

1910

1920

1930

1940

1950

1960

1970

1980

1990

2000

100 200 300Day of

Cal Year

Colorado River at Lees Ferry, AZ

100,000

10,000

1,000

Flow (cfs)

-8-7-6-5-4-3-2-1012345678

PDSI

Cal

enda

r Yea

r

Scales

SummarySummaryRasterRaster--based approachbased approach

Greater visualizationGreater visualizationRaster hydrographRaster hydrograph

Analyze temporal streamflow changeAnalyze temporal streamflow changeGrid correlogramGrid correlogram

Verify if calibration dataset is more Verify if calibration dataset is more ““naturalnatural””New approach to identify temporal variabilityNew approach to identify temporal variabilityEnhance and replicate streamflow conditionsEnhance and replicate streamflow conditions

Raster-based Streamflow Analysis BackgroundStudy sitesRaster gridded time seriesColorado River exampleColorado River exampleAutocorrelationCorrelogramsCorrelogramsArtificial flow examplesArtificial flow examples, part 2ResultsResultsResultsResultsResultsResultsClimate applicationsClimate applicationsOther applicationsSummary