33
Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

Embed Size (px)

Citation preview

Page 1: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

1

Upper Mississippi River System Flow Frequency Study

Rolf Olsen

Institute for Water Resources

U.S. Army Corps of Engineers

Alexandria, Virginia

Page 2: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

2

Outline

Background Projections of future climate Trends in hydrologic record Alternative statistical models that do not

assume “stationarity” Results – 1% flood estimates Data Quality Act petition Conclusion

Page 3: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

3

Background

Update the discharge frequency relationships and water surface profiles on Upper Mississippi River, Illinois River, and Missouri River below Gavins Point dam.

After 1993 Mississippi River flood, communities were questioning their flood risk.

Page 4: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

4

Study Area

Page 5: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

5

New Methods

Used unregulated flows at gages for flood frequency distribution.

Flood control reservoir project impacts defined by developing regulated versus non-regulated relationships for discharges.

UNET unsteady flow program used to address hydraulic impacts including levee performance.

Considered effects of climate change and variability.

Page 6: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

6

Projections of Future Climate

Page 7: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

7

Projected Future Temperature

Page 8: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

8

Projected Future Precipitation

Page 9: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

9

Changes in Runoff based on GCM Simulations

 

Estimated Runoff and Changes (mm) Region Historical Canadian Hadley

1961-1990 2025-2034 2025-2034 Upper Mississippi 195 - 42 42 Missouri 40 - 9 7

Runoff estimated using simple mass-balance model and temperature and precipitation from GCMs

General Circulation Models (GCMs) do not agree on whether annual runoff will increase or decrease

Page 10: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

10

Conclusions: Future Climate

Results of General Circulation Models used to project future climate are still ambiguous.

Although flood magnitudes and frequencies may change as a result of global warming, evidence is not strong enough to project even the direction of change for the Upper Mississippi and Missouri River basins.

Page 11: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

11

Trends in Hydrologic Record

Page 12: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

12

Evidence of increasing temperature and precipitation in the region.

Evidence flood risk may have changed over time for some stations in basin.

Conclusions: Trends

Page 13: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

13

Upper MississippiTrends

Annual Flood (1-Day High)

Annual Average Flows 1-Day Low Flows

Page 14: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

14

Significant Flood Trends

Keokuk

Illinois River

Hannibal

HermannSt. Louis

Trends on main stem using unimpaired flows

Page 15: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

15

Trends in Annual Flood

Illinois River at Meredosia

0

20000

40000

60000

80000

100000

120000

140000

1921

1926

1931

1936

1941

1946

1951

1956

1961

1966

1971

1976

1981

1986

Year

Dis

char

ge

(cfs

)

Page 16: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

16

Comparison of Palmer Drought Severity Index Based on Tree Rings and Annual Average Flow of Mississippi River at Keokuk, Iowa (10-Year Moving Averages)

-1.50

-1.25

-1.00

-0.75

-0.50

-0.25

0.00

0.25

0.50

0.75

1.00

1.25

1.50

1705

1725

1745

1765

1785

1805

1825

1845

1865

1885

1905

1925

1945

1965

1985

Year

Pal

mer

Dro

ug

ht

Sev

erit

y In

dex

-1.50

-1.25

-1.00

-0.75

-0.50

-0.25

0.00

0.25

0.50

0.75

1.00

1.25

1.50

Mis

siss

ipp

pi

Riv

er a

t K

eoku

k (N

orm

aliz

ed F

low

)

Tree Ring PDSI

Annual Flow at Keokuk

Paleoclimate

Page 17: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

17

Alternative Statistical Models(that do not assume “stationarity”)

Page 18: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

18

Alternative Models -1

Incorporate trend into statistical model– Trend analysis depends on the period of time

used in the analysis. – Requires a subjective evaluation of when the

trend begins or ends.– Uncertain how to extrapolate the trend beyond

the period of record.

Page 19: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

19

Alternative Models – 2

Employ selected subset of historical years to represent flood frequency distribution. (Use more recent period in place of the entire period of record).– If a short period of record is used, the results will lack

precision. – Problematic to select representative subset of record. – Although it may be possible to determine “climate

regime” for past periods of time, currently unable to predict when regime will shift.

Page 20: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

20

Alternative Models – 3

Condition flood series on climate indices (flood record is derived from low-frequency climate variations and climate indices represent the underlying climatic conditions).– Lack skill in predicting decadal climatic

fluctuations. – Can assume low frequency index persist over

next year, but planning generally requires longer planning horizon.

Page 21: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

21

Alternative Models – 4

No longer assume that the random variables are independent over time (pattern of episodic wet and dry periods that persist over several years). – Resulting variation in flood risk is unlikely to

affect flood risk management.– Larger standard errors of mean and 100-year

flood estimators than a model assuming independence.

Page 22: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

22

Flood sequences affected by trends and interdecadal climate variability may be described as realizations of stationary persistent processes.

Stationary time series allow risk to vary over time but preserve the assumption that hydrology is stationary in the long run.

Conclusions: Alternative Models

Page 23: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

23

Recommendations

There is not enough compelling evidence to deviate from application of log-Pearson III distribution estimated by application of the method of moments to log flows.

Currently no viable alternative in flood frequency analysis to using the assumption that flood flows are independent and identically distributed random variables.

Page 24: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

24

Results

Page 25: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

25

Comparison of 1% Chance Floods – Missouri River

Location 1962 Study 2003 Study Difference

St. Joseph, MO

815.1 819.4 4.3

USGS Gage 270,000 cfs 261,000 cfs

Kansas City, MO

748.5 749.5 1.0

USGS Gage 425,000 cfs 401,000 cfs

Hermann, MO

518.4 518.6 0.2

USGS Gage 620,000 cfs 673,000 cfs

Page 26: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

26

Comparison of 1% Chance Floods – Mississippi River

Location 1979 Study 2003 Study Difference

Keokuk, IA 500.8 501 0.2

USGS Gage 351,000 cfs 366,000 cfs

Hannibal, MO

475.3 477.1 1.8

374,000 cfs 440,000 cfs

St. Louis, MO

427 426 -1

USGS Gage 1,020,000 cfs

910,000 cfs

Page 27: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

27

Data Quality Act Petition

Page 28: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

28

Data Quality Act Petition

Act requires Federal agencies to “issue guidelines ensuring and maximizing the quality, objectivity, utility, and integrity of information (including statistical information)” and allow “affected persons to seek and obtain correction of information.”

Submitted by Missouri Coalition for the Environment Foundation

Page 29: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

29

Data Quality Act Petition

“The flow frequency study relies on a key assumption that flooding has been stationary over the period of record, an assumption that lacks factual support.”

“The existence of a trend of increasing flooding on the Midwest’s large rivers is also supported by a growing body of scientific literature, but the Corps almost completely ignored such literature when conducting the flow frequency study.”

Page 30: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

30

Data Quality Act Petition

“Report singles out two papers, Wolock and McCabe (1999) and Lins and Slack (1999) and uses these references to conclude that both studies of past flood trends and Global Climate Model (GCM) simulations of future changes in flood occurrence yield ambiguous and uncertain results.”

Page 31: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

31

Conclusion

Page 32: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

32

Under conditions of climate change and variability, the uncertainty in the estimate of the 1% flood increases.

Uncertainty in flood risk estimates should be communicated to floodplain communities and local sponsors of flood control projects.

National Flood Insurance Program delineation of Special Flood Hazard Areas as 1% floodplain appears arbitrary.

Conclusion

Page 33: 1 Upper Mississippi River System Flow Frequency Study Rolf Olsen Institute for Water Resources U.S. Army Corps of Engineers Alexandria, Virginia

33

No uniform agreement on how to model non-stationarity.

Flood frequency analysis with climate change could become more subjective and “political.”

Federal agencies should consider updating Bulletin 17-B with one topic being how to treat interdecadal climate variability in flood risk assessment.

Conclusion