Silvia Trini Castelli - ISAC CNR

1. II esperimento Jack Rabbit:
confronto pieno di ostacoli
tra approcci modellistici
Bertrand Carissimo (CEREA), Silvia Trini Castelli (ISAC-CNR),
Gianni Tinarelli (ARIANET), Steve Hanna (Hanna Consultants),
Michael Pirhalla, Steven Perry, David Heist (US EPA)
Joseph Chang (RAND Corporation)

2. Special Sonic Anemometer Study in Fall 2015
and Spring 2016 around CONEXs
• Sonic anemometers were not used during JR II in the CONEX array because they would be
destroyed by the presence of chlorine
• Still, we want to know flow and turbulence in wakes next to CONEXs
• Special studies were carried out during about 40 days in fall 2015 and spring 2016 with
CONEXs in place, and no chlorine released
• 17 sonic anemometers were placed near the 2 by 3 CONEX stack and 13 near 40 ft
CONEX. Sonic “towers” were located upwind and downwind of obstacles.
32 m tower

3. 2m
8m
4m
16m
32m
32m Upwind Sonic Tower
• In addition to the sonics in the
CONEX array, there was a 32m
tower that provided an
unobstructed upwind flow profile at
5 levels
• 30-minute average mean and
turbulence variables were
calculated from the raw 10Hz sonic
data
32m tall tower 100m upwind from
CONEX array

4. Objectives
• JRII-Sonic (JRIIS) data processing and analysis
• Use JRII-Sonic for model validation, intercomparison
and improvement
• Validation of complex models with detailed sonic
measurements
• Intercomparison with simpler operational models
• Improvements of all models
• Test impact of wake prediction and differences on
the simulated concentration (no measurements)

5. 5
CONEX
Height (m)
1m AGL
Tower 1, 2, 5m
Tower 1, 2, 5, 10m
Sonic Observations
1m mast on CONEX
top
S111
S81 S83
S101
S82L1,2,3,4
S61L1,2,3,4
S73
S23
S72
S75
S74
S71
S51
S41
S24
S21
S22
S31L1,2,3
S11L1,2,3
S25
*Inner
subset
of the
array
around
sonics*
C1-C6
additional
model
sources
C1 C2 C3
C4 C5 C6

6. JRIIS Data Analysis
• Time periods informed by 32m tower were
selected for further analysis for sonics within array
• Filtered 30-min periods with relatively moderate
winds speeds characteristic of more neutral PBL
flow
• ≥ 2.5 m/s
• ±15º perpendicular to CONEX face
• selection for further analysis of wakes and
recirculation zones
Southerly flow:
Measuring period: March 24, 0730-2100
Selection: 2 cases at 11h00 and 13h30 (30 sonics working)
Choice :

7. Eulerian and Lagrangian approaches
• Mean advection-diffusion equation for a scalar c:
• Velocity and turbulence fields  solved by the CFD code
Code_Saturne using RANS models with classical k-ε or Rij-ε
closures adapted to the atmosphere and complex geometries
• Langevin equation for the particle’s motion:
EULERIAN APPROACH LAGRANGIAN APPROACH
 tdWuxbdtuxatdu ),(),()(   dttudx
 


t
V
tdrdtrStrtrPtrC

),(),|,(),(
V = entire volume of the fluid; P = probability density
function; S = source distribution of the contaminant
    dttdWtdW  2
;0
deterministic term stochastic term
incremental Wiener process:
dt)(a ux, )t(dW)(b ux,
dW

8. A few details:
- Two periods on 24/3 : 11h00 and 13h30
- Here : comparisons on cross sections at z=1m
- Work in progress : pointwise comparisons and statistics with
the sonics (velocity and turbulence)
- Model/model comparison for concentration
3 models so far in intercomparison:
• Code_Saturne + Mixing Length (Lm=1 and 2 m) (Eulerian)
• Code_Saturne + k-eps (Eulerian)
• PMSS : Micro Swift Spray (Eulerian diagnostic+Lagrangian)
Intercomparison

9. Micro-SWIFT SPRAYCode_Saturne , k-eps
Code_Saturne , Mixing length Lm=1m
Wake comparisons 24/3 at 11h00: Wind velocity
Code_Saturne , Mixing length Lm=2m

10. Micro-SWIFT SPRAYCode_Saturne , k-eps
Code_Saturne , Mixing length Lm=1mCode_Saturne , Mixing length Lm=2m
Wake comparisons 24/3 at 11h00: Source 2- conc

11. Micro-SWIFT SPRAYCode_Saturne , k-eps
Code_Saturne , Mixing length Lm=1mCode_Saturne , Mixing length Lm=2m
Wake comparisons 24/3 at 11h00: Source 5- conc

12. Micro-SWIFT SPRAYCode_Saturne , k-eps
Code_Saturne , Mixing length Lm=1m
Lm2CS_Vel
Code_Saturne , Mixing length Lm=2m
Wake comparisons 24/3 at 13h30: Wind velocity

13. Micro-SWIFT SPRAY
Code_Saturne , Mixing length Lm=1m
Code_Saturne , k-eps
Code_Saturne , Mixing length Lm=2m
Wake comparisons 24/3 at 13h30: Source 2- conc

14. Micro-SWIFT SPRAY
Code_Saturne , Mixing length Lm=1mCode_Saturne , Mixing length Lm=2m
Code_Saturne , k-eps
Wake comparisons 24/3 at 13h30: Source 5- conc

15. Conclusions
• Ongoing analysis of JRII-Sonic dataset
• Simulations for 2 cases 24/3 at 11h00 and 13h30
• 6 (virtual) passive sources added to test model sensitivity to wake
modeling
• Important sensitivity to type of wake modeling and turbulence closure
Work in progress
• Pointwise comparison statistics on velocity, turbulence with sonic
measurements and with the different models
• Diagnose model shortcomings and improve
• Comparison with EPA and U of Arkansas laboratory scale releases
• Invitation to other modelers to join

16. Digressione: NOSE e la modellistica!
NOSE
Network for Odour
Sensitivity
Sistema di
segnalazione delle
emissioni odorigene

17. NOSE
Network for Odour Sensitivity
Sistema di segnalazione delle emissioni odorigene
Consente ai cittadini di inviare le proprie segnalazioni georeferenziate
in tempo reale ed in modo riservato ad ARPA, riguardo la
localizzazione delle molestie olfattive avvertite nei Comuni dell’area.
WEB-APP gratuita ed anonima, utilizzabile da smartphone
nose-cnr.arpa.sicilia.it
Durante la segnalazione dell’evento odorigeno, attraverso una scelta multipla, il
cittadino deve fornire informazioni riguardanti i seguenti parametri:
(i) tipo di odore percepito
(ii) intensità dell’odore
(iii) tipo di disturbo fisico rilevato
Sono inoltre a disposizione un certo numero di caratteri per eventuali descrizioni
integrative e commenti.
P.Bonasoni, S.Gilardoni, P.Malguzzi, T.Landi, S.T.Castelli, O.Drofa, G.Resci*, E.D’Accardi°, V.Infantino°

18. Una NOSE-ROOM per ARPA, Sindaci,
Protezione civile, Prefettura, CNR, …
raccoglie segnalazioni in NRT, elabora
le info ed in caso di necessità attiva
un apposito sistema di ALERT (vs
referenti, campionatori, …)
276
segnalazioni
EVENTO 19 Settembre 2019
Augusta 19/9/19
analisi chimica canister
Propilmercaptano:
118.1 µg/m3 (s.o. 0.2 µg/m3)
Butilmercaptano:
42.7 µg/m3 (s.o. 2.0 µg/m3)
Segnalazioni
orarie (10’)
I colori verde, giallo, arancio,
rosso descrivono il numero di
segnalazioni crescenti.
Tipologia di odore percepito
a seguito delle segnalazioni
ricevute: 220 idrocarburi, 22
zolfo, 19 altro, 10 solventi, 4 bruciato,
1 fognatura.

19. NOSE - Network for Odour Sensitivity 2019-2021
Progetto per lo sviluppo di attività di ricerca scientifica con finalità operative
al fine di caratterizzare ed identificare le aree sorgenti dei miasmi olfattivi nei
territori AERCA di Siracusa, Melilli e Gela, con applicabilità ad altre aree di interesse
• Analisi numeriche pluriennali di traccianti odorigeni (NMHC, H2S, Benzene, …)
• Preparazione, realizzazione e messa a punto APP NOSE (gestione privacy, sicurezza
dei dati, aspetti epidemiologici, …)
• Interfacciamento APP con Sistema campionamento automatico
• Analisi chimiche ed olfattometriche campioni (ARPA, Università CT)
• Analisi numeriche
• Messa a punto tool backtrajectories:
da MOLOCH (traiettorie deterministiche) a SMART (suite meteo-dispersiva)
• Interfacciamento APP con tool backtrajectories in NRT
• Chiusura suite NOSE operativa: APP-chimica-analisi-backtrajectory
 Connessione NOSE - epidemiologia
 Misure IEMS indoor/outdoor scuola

20. Back-trajectories:
da MOLOCH (traiettorie deterministiche) a SMART (suite meteo-dispersiva)
NOSE
19.09.2019
Spray-Moloch Atmospheric Regional Tool