The ECMWF Re-Analysis of the MAP SOP: final report available Christian Keil, ECMWF, Reading, England.
A re-analysis of the Mesoscale Alpine Programme (MAP) Special Observing Period (7 September to 15 November 1999) has been performed at ECMWF. Substantial upgrades in the global data assimilation system 4D-Var (T511/159L60, horizontal resolution approximately 40km) combined with special MAP observations provide a new reference description for studies on mountain related atmospheric phenomena. In the Alpine region, three times more humidity observations from radiosondes and five times more humidity observations at surface stations than in operations~1999 have been assimilated. Four out of 16 European windprofilers were not used because of variable quality. A control analysis excluding the MAP special observations is available, too. Some key results:
1. Assimilating MAP observations slightly moistens the troposphere in the southern Alpine region, southern France and over the Adriatic Sea. Conversely, European windprofilers led to slightly drier conditions south of the Alps.
2. Windprofilers reinforce dynamical features as shear zones and jet-stream winds (IOP-15). However, windprofilers operated at different microwave frequencies (detecting different vertical ranges) don't report necessarily consistent information when deployed at the same location (e.g. VHF and UHF Lonate). Here, a cross-check of the observed overlapping data is necessary prior to its assimilation.
3. A case-study-type investigation of IOP-2a shows that windprofiler data may increase the subsidence and lead to a drier boundary layer. This is caused by an erroneous feed-back on the humidity field through variations in the humidity background standard deviation calculation (in 4D-Var) and not through the model dynamics.
4. The remarkable agreement of rainfall time series in the Po catchment highlights the model skill. The model is able to capture the different precipitation patterns, fingerprints of different flow regimes, south of the Alps, e.g., during IOP-2b and 8.
5. Validation with model-independent GPS derived integrated water vapour contents shows that the humidity field describes well the inter-diurnal humidity variations at southern Alpine GPS stations (Fig.1). However, the analysis seems to be slightly too dry compared with GPS measurements. For instance, the mean analysis increments of north-Italian GPS stations depicted in Fig.1 amount to 2.9 kg/m2 at Genova and 0.5 kg/m2 at Medicina. Note that the dry bias is even larger for the operational analysis in 1999.
The following data have been transferred to the MAP Data Centre (MDC): (a) All uniformly formatted MAP-SOP observations (in BUFR format). (b) 3 hourly re-analyzed fields (variables Z, U, V, W, T, P) on model levels for the European region together with (c) analyzed fields of the control analysis using the same ECMWF model system excluding the additional MAP-SOP data.
This shall allow detailed investigations dealing with the MAP SOP, for instance, performing high-resolution model forecasts or mesoscale data assimilation experiments.
Keil, C. and C. Cardinali (2003): The ECMWF Re-Analysis of the Mesoscale Alpine Programme Special Observing Period. ECMWF Technical Memorandum 401. Available from ECMWF, Shinfield Park, Reading, RG2 9AX, UK.
