a process-based “bottom-up” approach for addressing changing flood-climate relationships

A Process-Based “Bottom-Up” Approach for Addressing Changing Flood-Climate

Relationships

COHS Workshop National Academy of Sciences

Global Change and Extreme Hydrology: Testing Conventional Wisdom

January 5-6, 2010Katie Hirschboeck

How can water managers deal with events in the “tails” of streamflow probability distributions, both floods

and droughts,

. . .by moving beyond conventional wisdom and

approaches . . . .

“GLOBAL CHANGE & EXTREME HYDROLOGY”

are planning for future extreme LOW FLOW conditions using :

-- tree-ring reconstructions -- simulations-- scenario-building-- climate projection

modeling

WESTERN WATER MANAGERS :

Water supply simulation based on extreme low flow sequences in the

paleo-record

STREAMFLOW RECONSTRUCTION for 1330-20052002 & 1996 = lowest annual flows in the entire record

In contrast . . .

FLOOD HAZARD MANAGERS:

have been more constrained in developing ways to incorporate climate change information operationally due to:

-- existing flood management policy and practices

-- the short-term, localized, and weather-based nature of the flooding process itself

Information presented in an operationally useful format for flood managers which describes how changes in the large-scale climatic “drivers” of hydrometeorological extremes will affect flooding variability in SPECIFIC WATERSHEDS

What’s Needed:

This presentation argues: . . . that attention to some very basic elements at the local and regional

watershed scale -- such as basin size, storm type,

seasonality, atmospheric circulation patterns, and geographic setting

. . . can provide a basis for a cross-scale approach to linking GLOBAL climate variability with LOCAL hydrologic variations . . .

. . . let the rivers “speak for themselves” about how they respond to climate !

In other words we will . . . .

Santa Cruz River at Tucson, Arizona

I. RE-EXAMINING CONVENTIONAL WISDOM & ASSUMPTIONS:The Standard iid Assumption for FFA

II.RE-THINKINGNew Insights from “Flood Hydroclimatology”

III.THE “BOTTOM-UP” APPROACHComplementary Upscaling

IV.FINAL THOUGHTS

OUTLINE

I. RE-EXAMINING

http://acwi.gov/hydrology/Frequency/B17bFAQ.html#mixed

“Flood magnitudes are determined by many factors, in unpredictable combinations.

It is conceptually useful to think of the various factors as "populations" and to think of each year's flood as being the result of random selection of a "population”, followed by random drawing of a particular flood magnitude from the selected population.”

“ iid ” assumption: independently, identically

distributed

The standard approach to

Flood Frequency

Analysis (FFA) assumes

stationarity in the time series

& “iid”

The Standard iid Assumption for FFA

Summer convective event

Synoptic-scale winter event

Tropical storm or other extreme event

The type of storm influences the shape of a hydrograph and the magnitude & persistence of the flood peak

This can vary with basin size (e.g. convective events are more important flood producers in small drainage basins)

Storm type hydrograph

The way in which rainfall is delivered

• in both space (e.g., storm movement, direction) • and time (e.g., rainfall rate, intensity)• over drainage basins of different sizes & orographies

In addition, extreme flow events can emerge from synergism in:

from Doswell et al. (1996)

Meteorological & climatological flood-producing

mechanisms operate at

varying temporal and spatial scales

FLOOD-CAUSING MECHANISMS

HYDROCLIMATOLOGY

Weather, short time scales Local / regional spatial scales Forecasts, real-time warnings

vs.

Seasonal / long-term perspective Site-specific and regional synthesis of

flood-causing weather scenarios Regional linkages/differences identified Entire flood history context

benchmarks for future events

HYDROMETEOROLOGY

It all started with a newspaper ad . . . .

Re-Examining the “iid” Assumption

THE FFA“FLOOD PROCESSOR”

With expanded feed tube – for entering all kinds of flood data

including steel chopping, slicing & grating blades

– for removing unique physical characteristics, climatic information, and outliersplus plastic mixing blade – to mix the populations together

Alternative Conceptual Framework:

Time-varying means

Time-varying variances

Both

SOURCE: Hirschboeck, 1988

Mixed frequency distributions may arise from:• storm types• synoptic patterns• ENSO, etc. teleconnections• multi-decadal circulation regimes

II. RE-THINKING

FLOOD HYDROCLIMATOLOGY is the analysis of flood events within the context of their history of variation - in magnitude, frequency, seasonality - over a relatively long period of time - analyzed within the spatial framework of changing combinations of meteorological causative mechanisms

SOURCE: Hirschboeck, 1988

Flood Hydroclimatology Approach

Meteorological / Mechanistic / Circulation-Linked

Flood Hydroclimatology Framework / Link to Flood Distribution

“ Bottom–Up ” Approach(surface-to-atmosphere)

Observed Gage Record

3 EXAMPLES: Flood Hydroclimatology in AZ

Sample Distributions of

Peaks-above-Base (Partial

Duration Series) events:

Are there climatically controlled mixed populations within?

Santa Cruz River at TucsonPeak flows separated into 3 hydroclimatic subgroups

Hirschboeck et .al. 2000

Tropical storm Sumer

Convective

Winter Synoptic

All Peaks

What does this time series look like when classified hydroclimatically?What kinds of storms produced the biggest floods?

Santa Cruz at Tucson

0

5000

10000

15000

20000

25000

30000

35000

40000

45000

50000

1915 1920 1925 1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000

Water Year

Dis

char

ge in

(cf

s)52700 (cfs)

Convective

Tropical Storm

Synoptic

Hydroclimatically classified time series . . .

Hirschboeck et .al. 2000

Verde River below Tangle Creek

Peak flows separated into 3 hydroclimatic subgroups

Tropical storm

Sumer Convectiv

e

Winter SynopticAll Peaks

Historical Flood

Annual floodpeaks only:

Empirical plotting positions computed separately for each hydroclimatic type

Sample frequency curve defined by plotting observed flood magnitudes vs their empirical probability plotting positions, separated by flood type

Alila & Mtiraoui 2002

Probability analysis based on

hydroclimatically separated

flood series

Thinking Beyond the Standard iid Assumption for FFA . . . .

Based on these results we can re-envision the underlying probability distribution function for Arizona floods to be not this . . . .

Alternative Model to Explain How Flood Magnitudes Vary over Time

Schematic for Arizona floods basedon different storm types

Varying mean and standard deviationsdue to different causal mechanisms

. . . but this:

HOW MIGHT CLIMATE CHANGE AFFECT THESE

DISTRIBUTIONS?

Some Important Flood-Generating Tropical Storms

Tropical storm Octave Oct 1983

Change in Frequency or Intensity of Tropical Storms?

Latitudinal Shifts inWinter Storm Track?

Roosevelt Dam Jan 1993 Winter flooding

on Rillito in Tucson

More Intense Summer Monsoon?

Sabino Creek July 2006

When the dominance of different types of flood-producing circulation patterns changes over time, the probability distributions of potential flooding at any given time (t) may be altered.

Conceptual Framework for Circulation Pattern Changes

El Nino year

La Nina year

Blocking Regime

Zonal Regime

. . . or this:

Conceptual Framework for Low-Frequency Variations and/or Regime Shifts:. . . or this:

A shift in circulation or SST regime (or anomalous persistence of a given regime) will lead to different theoretical frequency / probability distributions over time.

Hirschboeck 1988

Flood Hydroclimatology for Floods of Record

after Costa (1985)

Extreme Floods of Record evolved from: • uncommon (or unseasonable) locations of typical circulation features (a future manifestation of climate change?)• unusual combinations of atmospheric processes• rare configurations in circulation patterns (e.g. extreme blocking)• exceptional persistence of a specific circulation pattern.

Lane Canyon flash flood

EXAMPLE:

Rare configurations in circulation patterns (extreme blocking)

Jimmy Camp Creek flood of 1965

EXAMPLE:

exceptional persistence of a specific circulation pattern.

OVERALL:Unusually large floods in drainage basins of all sizes are likely to be associated with circulation anomalies involving quasi-stationary patterns such as blocking ridges and cutoff lows in the middle-level flow.

III. THE BOTTOM-UP APPROACH

Hirschboeck 2003 “Respecting the Drainage Divide” Water Resources Update UCOWR

Interpolation of GCM results computed atlarge spatial scale fields to higher resolution, smaller spatial scale fields, and eventuallyto watershed processes at the surface.

DOWNSCALING

“Scaling up from local data is as important as scaling down from globally forced regional models.”

— Pulwarty, 2003

PROPOSED COMPLEMENTARY APPROACH:

RATIONALE FOR PROCESS-SENSITIVE UPSCALING:

Attention to climatic driving forces & causes:

-- storm type seasonality-- atmospheric circulation patterns

with respect to:-- basin size -- watershed boundary / drainage divide

-- geographic setting (moisture sources, etc.). . . can provide a basis for a cross-scale linkage

of GLOBAL climate variability with LOCAL hydrologic variations

at the individual basin scale . . .

• Process-sensitive upscaling . . .can define relationships that may not be detected via precipitation downscaling

• Allows the imprint of a drainage basin’s characteristic mode of interacting with precipitation in a given storm type to be incorporated into the statistics of the flow event’s probability distribution as it is “scaled up” and linked to model output and /or a larger scale flow-generating circulation pattern

IV. FINAL THOUGHTS

Is this evidence of climate change?

Extreme events have a legacy of confounding us!

Overall Recommendation:A systematic compilation of watershed-specific information about spatially and temporally varying hydroclimatic extremes is proposed as a starting place for making operationally useful decisions about prospective climatic changes.

1. The impact of climate change on a flood distribution is likely to be more complex than a simple shift in mean or variance

2. Climatic changes can be conceptualized as time-varying atmospheric circulation regimes that generate a mix of shifting streamflow probability distributions over time

Recommendation: We need new and evolving statistical tools that can address this behavior.

3. The interactions between storm properties and drainage basin properties also play an important role in the occurrence and magnitude of large floods both regionally and seasonally.

Recommendation: Watershed–based hydrometeorology studies should be a key component of watershed and flood management practice.

4. Shifts in storm track locations and other anomalous circulation behavior are clearly linked to unusual flood (and drought ) behavior.They are likely to be the factors most directly responsible for projected increases in hydrologic extremes under a changing climate.

Recommendation: Use process-sensitive upscaling to link circulation patterns directly to flood–producing mechanisms and to complement downscaling