Assessing Volcanic Ash Hazard by Using the CALPUFF System

1
Assessing Volcanic Ash Hazard by Using the
CALPUFF System
The 2nd International Conference on Volcanic Ash
and Aviation SafetyJune 21-24, 2004Alexandria,
Virginia (USA)

• S.Barsotti, A.Neri, J.Scire
• Istituto Nazionale di Geofisica e Vulcanologia,
• Centro per la Modellistica Fisica e Pericolosità
  dei Processi Vulcanici, Pisa, Italy
• Earth Tech, Inc., Concord, Massachusetts, USA

2
Outline

• Motivation and objective
• The CALPUFF modeling system structure
  (geophysical pre-processors, meteo diagnostic
  code CALMET and the dispersion code CALPUFF)
• Results of a preliminary application of CALPUFF
  to the July 2001 Mt.Etnas eruption
• Conclusive remarks

3

• Motivation
• During the last eruptions of Mt.Etna (2001 and
  2002) the relatively minor explosive activity
  renewed the interest in the volcanic risk due to
  the presence of ash in the atmosphere

Objective
To develope a predictive modeling tool - able to
forecast the movement and properties of an ash
cloud in the atmosphere - to be used during a
volcanic crisis
4
Fig1 Picture of the city of Catania taken from
the seaport during the 2001 eruption
Fig3 The main air routes that connect the
Fontanarossa and Sigonella airports with national
and international destinations (figure provided
by Cap.G.A.r.s. Geof. Franco Colombo)
Photo by Franco Crisafi
Fig2 Photo of the ash cloud as seen from the
airport of Catania
5
CALPUFF Modeling System (Scire et al. 2000)

• Developed in the 90s by Sigma Research
  Corporation (now part of the Earth Tech, Inc.) as
  air quality modeling system
• Proposed by U.S. EPA as a Guideline model for
  regulatory applications

System Flowchart
Meteo data (MM5 or CALETA)
Terrain elevation data (TERREL)
Landuse data (CTGPROC)
CALMET(diagnostic meteo processor)
Emission data
CALPUFF (Lagrangian dispersion code)
6
CALMET (diagnostic meteo processor)

• Input data P, Elevation m.s.l, T, wind
  direction, wind speed, vertical velocity,
  relative humidity at different levels and every
  three hours
• Output data U, V, W wind components, T and all
  the micrometeorological variables on a specified
  grid and every hour

HOW DOES IT WORK?
Prognostic code output
Initial guess field

• Refinement by interpolation
• Correction by orography
• Divergence minimization

Terrain effects
Step 1 wind field

• Weighted observed data
• Divergence minimization

Objective analysis
Step 2 wind field
7
CALPUFF

• It is a Lagrangian dispersion model that treats
  the emitted material as a sequence of packets, or
  puffs, containing discrete quantity of particles.
  Each puff is then described with a Gaussian
  distribution and subjected to advection,
  diffusion and gravity
• Selected a set of receptors, gridded or discrete,
  in the domain investigated, CALPUFF computes
  hourly the concentration and flux of particles at
  these points
• Main features

• Different granulometric classes
• Time dependent emission and meteo data
• Dry and wet deposition
• Buoyant area source
• Domain from tens of ms to hundreds of kms
• from the source

8
The basic equation
is the material mass in the puff and is a
function of the sampling interval. Its variations
are due to removal and to chemical processes.
The sigmas are the horizontal Gaussian dispersion
coefficients and are functions of the sampling
interval. They include effects of atmospheric
turbulence, plume buoyancy and lateral
(crosswind) scale of an area-source.
is the vertical term of the Gaussian
distribution. It takes account for the vertical
diffusion due to the mixing lid and plume
buoyancy and is a function of the distance from
the source.
9
Turbulence and deposition modeling

• The atmospheric turbulence coefficients can be
  calculated on different meteo data
• measured values of turbulence,
• micrometeorological variables calculated by
  CALMET,
• PG dispersion coefficients calculated on the
  values of the stability classes provided by
  CALMET.

• The dry deposition model is based on an approach
  which expresses the deposition velocity as the
  inverse of a sum of resistances (opposing to
  the transport through the atmosphere) plus
  gravitational settling terms (using Stokes
  equation).

10
Area Source Plume Rise

• One of the main features of CALPUFF is the
  capability of treating the case of very hot area
  source. Using a file containing all the
  information related to the source, the code
  solves the three fundamental equations of
  conservation of mass, momentum and energy.

The equation system is (Weil 1988)
11
A First Volcanological Application(Etna July
2001)

• Simulation parameters
• Computational domain
• Cell numbers 100X100
• ?grid2km
• Grid origin xUTM(km) 411.021
• yUTM(km) 4095.136
• Period
• 20-24 July 2001

• Geo and meteo input data
• GTOPO Global data 900m res.
• USGS Global land-use data set 900m res.
• ECMWF meteo data 2.52.5 degree res.
• NOAA meteo data 2.52.5 degree res.

12
Source data
Coltelli and Macedonio, GNV Ass. 2001

• For the rising plume
• Exit velocity 25-75 m/sec
• Exit temperature 100-300 C
• Initial radius 25-75 m
• Ground elevation 2550 m
• Emission height a.g. 100 m

Photo by Boris Bencke

• For the puff description
• Particles diameter 3-64 µm
• Particles density 2500 kg/m3
• Emission rate 102-103 kg/sec

Photo by Tom Pfeiffer
13
General features of ash dispersal
Ash concentration on air (Kg/m3)
14
Temporal Evolution of the Emission Point

• The code computes the effective height and the
  downwind distance from the vent where a puff
  description of the emission begins

15
Ash Cloud Concentration (1)Vertical distribution
of ash concentration along the dispersal axis

• Vertical distribution of particle concentration
  at a fixed time and different distances from the
  vent

• Temporal evolution of the vertical concentration
  at 25 km from the vent

16
The sensitivity to meteo data

• Image acquired by SeaStar Satellite at 1100UTC on
  20July 2001

• NOAA

• ECMWF

• Concentration (kg/m3) on air summed on different
  vertical levels

• Vertical distribution of wind direction

17
The sensitivity to meteo data

• Image acquired by SeaStar Satellite at 1129UTC on
  22July 2001

• NOAA

• ECMWF

• Concentration (kg/m3) on air summed on different
  vertical levels

• Vertical distribution of wind direction

18
Conclusive Remarks

• CALPUFF is a promising system for real-time
  forecast of volcanic ash dispersal produced by
  weak explosive activity
• High resolution meteo data and accurate
  description of plume rise are critical for a
  correct modeling of plume dispersal
• Definition of well-known test cases is
  fundamental for the validation of the code with
  observed data
• Real-time measurements of the eruption plume and
  dispersal are crucial in order to obtain reliable
  forecasts

19
(No Transcript)
20
CALPUFF Loop Structure
Begin HOUR Loop
Puff initialization
Wind initialization for the advection
Begin PUFF Loop
Begin SAMPLING Loop
End SAMPLING Loop
End PUFF Loop
Computes concentration at the receptors
End HOUR Loop
21
Ash Cloud Concentration (2) Horizontal
distribution of concentration along the dispersal
axis

• Ash concentration at 3500m above sea level

• Cumulative concentration on all vertical levels

22
The sensitivity to source data

• To investigate the effects of the variation of
  emission parameters on the rising plume and to
  take account of the error affects the measures,
  we have assumed exit velocity and temperature to
  be affected by an 50 error

• Variation on exit velocity

• Variation on exit temperature