Work Package Leaders
Lars-Anders Breivik (met.no) and Trond Iversen (met.no)
Other Key Investigators
Harald Schyberg, Inger-Lise Frogner, Øyvind Saetra
International partners
Dale Barker, Martin Leutbecher, Jan Barkmeijer
Background
New satellite instrumentation on operational meteorological satellites provides improved information on the Arctic atmosphere. We will utilize IPY campaign data with extended radiosonde observations for collocation with satellite data, to develop an advanced Arctic high-resolution atmospheric data assimilation system. This will be a part of international co-operation (HIRLAM, ECMWF, EUMETSAT) and other IPY projects.Good results have been obtained by assimilation of satellite microwave temperature profile data from AMSU-A (Advanced Microwave Sounding Unit) over olar areas. In periods with dominant mid-troposphere winds from the sea ice, added MSU-A observations over sea ice have been shown to have a positive impact on forecast quality for the Barents Sea, the Norwegian Sea and the Scandinavian peninsula (Thyess et al., 2005). We will focus on a new satellite sounding sensor which will become operational from March 2007: IASI (Infrared Atmospheric Sounding Interferometer) on the European polar satellite METOP, successfully launched in October 2006. Data from a similar American instrument (AIRS) have been shown to add significantly to the forecast quality in observation-sparse areas (Atlas, 2005). These instruments provide better information for the vertical profile and higher horizontal resolution than the AMSU instrument.
An ensemble prediction system (EPS) for the European-Atlantic sector of the Arctic is needed to enhance the predictive capability of rare and extreme weather events. Such a system for THORPEX-IPY would employ a high-resolution limited area model embedded into a global EPS. A few similar LAMEPS in other regions have been reported in later years (Marsigli et al., 2001; Frogner and Iversen, 2002; Frogner et al., 2006), and a HIRLAM-based LAMEPS has been run operationally at met.no since February 2005. The work proposed here will be part of a larger upcoming co-operation in the HIRLAM-A project.
Research plan
With respect to the new satellite data, we will assess how best to exploit the observational information content in NWP models. This includes how to optimally reduce the large data volume related to the large number of sounding channels. Furthermore, we will assess the cloud screening, in particular over Arctic snow and sea ice surfaces. This is because clouds have a strong influence on the infrared channels used in IASI, and current cloud screening is not optimal for use over snow and ice. The work will be carried out in co-operation with the HIRLAM data assimilation group and the OSI (Ocean and Sea Ice) SAF. We will fit the existing HIRLAM-based LAMEPS to a target domain which includes the European-Atlantic sector of the Arctic. The system will estimate forecast spread from uncertain initial and lateral boundary data based on the operational ensemble system run at ECMWF, presently at resolution T399L62. It is planned to run LAMEPS with 0.2 degree or finer resolution, depending on resources. A 20-member ensemble, in addition to a control run based on the HIRLAM data assimilation, is planned for. We will investigate the possibility and benefits of using a model version which calculates ocean waves on-line with the meteorology. Through HIRLAM and co-operation with KNMI, we will furthermore investigate the option of using internally developed optimal initial perturbations in the domain. A primary focus will be on polar lows and other extreme and rare wintertime events. A regular operational routine for the system will be set up during IPY. Improved model physics, data assimilation and options for creating ensembles will be tested in hindcast mode. We will use standard verification methods for probabilistic forecasts and standard measures for spread-skill relations (Saetra and Bidlot, 2004; Saetra et al., 2004, Frogner et al., 2006). The verification will make use of, and investigate the sensitivity to, the planned increased level of regular and satellite observations during IPY. A baseline for comparing the forecast quality will be the standard EPS run at ECMWF. The deliverables from this activity will be probabilistic forecasts up to 60 hours, together with scientific experiments and innovative developments. We will in particular investigate how observations for IPY and parameterizations developed in WP2 improve the probabilistic predictions.