3° ICTP Workshop “The Theory and Use of Regional Climate Models” May 29 – June 9

transcript

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

CETEMPS, Physics Department, University of L‘Aquila

http://cetemps.aquila.infn.it

Marco Verdecchia

CHyM: an operational distributed hydrological model

using different data sources

3° ICTP Workshop on “The Theory and Use of

Regional Climate Models”May 29 – June 9 2006

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

Developed by:

Since 2002

•It runs in any geographical domain with any resolution•It runs in any (?) unix platforms•It does not use any licensed software•Different kind of rain data are assimilated

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

Two reasons for NOT using CHYM:

• NOT easy to use (CHyM is a NSE2USE Model)

• Documentation is not (yet) available

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

Outline• Description of the model • Algorithms Generating streamflow network Building prec. fields with diff. data• Physical processes• Applications

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

Why to develop a new Hydrological Model?

•It has been thought for operational purposes

•It is a good “exercise”

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

Step 1: generating streamflow network from DEM

DEM matrix for the selected domain and resolution is generated

Flow direction matrix is computed

Validation

“Pits” and singularities are corrected

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

DEM is available with a resolution of 300 m

For each cell the slope is computed as:

Runoff direction

Max slope

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

CHyM: Drainage network test

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

A Cellular Automata definitionA Cellular Automata definition

A cellular automaton is a A cellular automaton is a discrete dynamical discrete dynamical systemsystem Space, time and states of the system are discrete Space, time and states of the system are discrete quantitiesquantities

Each point in a regular spatial lattice, called a cell, Each point in a regular spatial lattice, called a cell, can have anyone of a finite number of statescan have anyone of a finite number of states

The state of the cells in the lattice are updated The state of the cells in the lattice are updated according to a according to a local rulelocal rule

All cells on the lattice are All cells on the lattice are updated synchronouslyupdated synchronously

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

A Cellular Automata A Cellular Automata applicationapplication

Life rules by Chris G. Langton

The status of each CA can be ON or OFF

If more than 3 CA in the neighborhood are ON CA became OFF

If less than 2 CA in the neighborhood are ON, CA became OFF

Otherwise CA became ON

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

CA for CHyM applicationsCA for CHyM applications

CHyM grid is considered an aggregate of cellular CHyM grid is considered an aggregate of cellular automataautomata

The status of a cell corresponds to the value of aThe status of a cell corresponds to the value of a

The state of the cells in the lattice is updated The state of the cells in the lattice is updated according to followingaccording to following rule rule

All cells on the lattice are updated synchronouslyAll cells on the lattice are updated synchronously

Update ends when flow scheme is OKUpdate ends when flow scheme is OK

8

)(j

ijjii hhhh

CHyM matrix (DEM)CHyM matrix (DEM)

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

CHyM: Recipe for DEM pit correction

•Smooth DEM using CA rules until FD can be obtained for all the cells

•Generate streamflow network using smoothed DEM

•Use “true” DEM and modify ONLY the cells draining toward an higher cell

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

CHyM: DEM pit correction

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

CHyM: Examples of Drainage Network Extraction

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

CHyM: the Rolling Stones Algorithm (RSA)

2. Each time that the stone goes through one cell for this cell a counter is incremented by 1

1. Starting from each cell a stone rolls up to the river‘s mouth

3. If a quantity A is associated to each stone where Ais equivalent to the surface where the stone was at the beginning, for each cell it can be computed theupstream drained surface

N

iiA

1

4. If a quantity R is associated to each stone where Ris equivalent to the precipitation where the stone was at the beginning, for each cell it can be computed theupstream drained precipitation

N

iiR

1

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

CHyM: the Rolling Stones Algorithm (RSA)

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

MuseoArchive

MM5 analysis MM5 Forecast

Now

timeRadar Gauge data

Satellite estimations

Step 2: Building Precipitation Fields using different Data Sources

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

CHyM rain data sources: Gauge measurements

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

CHyM rain data sources: Radar estimates

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

NERETIR

NEural RainfallEstimationfromTermalInfraRed

CHyM rain data sources: NERETIR

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

MM5 Meteorological Model

CHyM rain data sources: MM5

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

Step 2: Building Precipitation Fields using different Data Sources

Module 1

Module 2

Module 3

Module n

•Define subdomain•Fill cells corresponding to rain gauges•Fill subdomain matrix – Cr. Formula•Smooth subdomain matrix using CA algorithm

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

CA for CHyM applicationsCA for CHyM applications

CHyM grid is considered an aggregate of cellular CHyM grid is considered an aggregate of cellular automataautomata The status of a cell corresponds to the value of a CHyM matrix The status of a cell corresponds to the value of a CHyM matrix

The state of the cells in the lattice is updated The state of the cells in the lattice is updated according to followingaccording to following rule rule

All cells on the lattice are updated synchronouslyAll cells on the lattice are updated synchronously

Update ends when flow scheme is OKUpdate ends when flow scheme is OK

8

)(j

ijjii hhhh

But cells corresponding to rain gauges or defined in a previous Module are not updated

Update ends when a stable state is reachedUpdate ends when a stable state is reached

(RAIN)(DEM)

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

Observed rain using Cressman

algorithm

+ smoothing usingCellular Automata

jj ij

iji R

ar

arR

22

1

1rij≤ a

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

+ MUSEO module using Cressman

+ CA

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9In sequence …

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

CHyM Rain field sources: an example

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

Runoff EvapotraspirationInfiltrationRainfall

For each cell the simulated processes are:

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

CHyM: Infiltration

Soil moisture storage

Deep percolation

Infiltration

Interflow

Surface Runoff

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

CHyM: Evapotraspiration

Thornthwaite Formula (Thornthwaite and Mather, 1995)

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9CHyM: Runoff

Continuity equation

Momentum equation

cqx

Q

t

A

An

RSQ

3221

A= cross sectional area of the riverQ= flow rate of water dischargeqc= rain for length unit

S= slope

1/R= wetter perimeter

n= Manning‘s roughness coefficient

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

AIA

R

N

ii

N

ii

1

1 Ri = rain

Ai = drained surface

AI= Alarm Index

CHyM: A first application

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

Soverato Floodsimulation

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

Val Canale floodsimulation

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

CHyM: simulation of Aug 22-23 2005 event

Flood alert mapping using MM5 and CHyM

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

Flood alert mapping using MM5 and CHyM

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

Landslides prediction using CHyM. (Very) Preliminary results

•Daily precipitation from 1958 to 2002

•Total drained rain in the last N days

•Landslides occur when TDR-N is greater than LT (N and LT values to optimized)

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

CHyM: hydrological effects due to Alpine Glaciers Melting

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

For further information about the model and related literature please visit the URL:

http://cetemps.aquila.infn.it/chym/

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

CHyM: an operational distributed

hydrological model using different data sources

3° I

CT

P W

orks

hop

“

The

ory

and

Use

of

Reg

ion

al C

limat

e M

odel

s”M

ay 2

9 –

Jun

e 9

3° ICTP Workshop “The Theory and Use of Regional Climate Models” May 29 – June 9

Documents