Operational Dust Prediction - Mineral Dust: A Key Player in the Earth System - page 238

10.3.1 The International Cooperative for Aerosol Prediction

(ICAP) Multi-model Ensemble

As a result of the maturity of an international community of global aerosol forecast
model developers (Reid et al. 2011 ; Benedetti et al. 2011 ), the creation of broadly
acceptable norms, benchmarks and scorecards to evaluate aerosol forecast skill has
become an important issue. At the same time, the NWP community has recognised
the value in multi-model ensembles in developing probabilistic forecast tools.
Similarly, ensembles of global aerosol analyses are becoming an important tool for
climate studies (Huneeus et al. 2011 ). In response to community needs and views,
member developers of the International Cooperative for Aerosol Prediction (ICAP)
created a developmental global multi-model ensemble (MME) (Sessions et al.
2014 ) to allow exploration of relative differences between models and devise tools
for probabilistic prediction. Current models in the ICAP-MME dust component
include (1) NMMB/BSC-CTM, (2) ECMWF MACC, (3) JMA MASINGAR, (4)
NASA GEOS-5, (5) NOAA NGAC, (6) NRL developmental NAAPS and (7) NRL
20-member ensemble mean E-NAAPS (see Sect. 10.2.1 ). To allow for the inclusion
of quasi-operational models, the ICAP-MME is run 24 h behind actual time, which
reduces the forecast range from 5 to 4 days. Daily products include a host of
maps, mean-spread plots, verification plots and threat scores. As an example of
available products, Fig. 10.2 shows the 24 h dust forecasts of aerosol optical depth
at 550 nm wavelength valid for 2 January 2013, from the participating models. The
maps of AOD show a high degree of consistency of the dust forecasts among the
various models. However, there are differences in the magnitude of the AOD field
and also in some features, such as the plume being transported towards the North
Atlantic which is not present in all models. While currently ICAP-MME data is
only available to participating member centres, it is expected to be made public on
a quasi-operational basis by the end of 2014.

10.3.2 WMO SDS Regional Dust Prediction Multi-model

Ensemble

The Sand and Dust Storm Warning Advisory and Assessment System (SDS-
WAS) is a programme of the World Meteorological Organization (WMO) with the
mission to enhance the ability of countries to produce and deliver to end users
timely and precise sand and dust storm forecasts (Terradellas et al. 2011 ). The
WMO SDS-WAS Regional Centre for Northern Africa, Middle East and Europe
(NA-ME-E) coordinates the exchange of forecast products generated by different
dust models (BSC-DREAM8b v2, MACC, DREAM8-NMME-MACC, CHIMERE,
NMMB/BSC-Dust, MetUM, GEOS-5 and NGAC) and conducts a model inter-
comparison and evaluation within its geographic scope (Terradellas et al. 2012 ).
Two products describing centrality (multi-model median and mean) and two