Figure 1-1. The Alps. Scale changes at 200, 500, 1000, 1500, 2000, 2500, and 3000m altitude. Terrain above 3000m in white.
The complex topography of the earth's continents acts as a strong and permanent modifier of atmospheric circulations on a wide variety of scales. Certain mountain ranges such as the Alps exert a particularly large impact on the weather and climate of their environment.
The Alps have a length of ~800km, a width of ~100km, and are characterized by numerous different valley systems (Fig. 1-1). Maximum peak height reaches up to ~4800m. Atmospheric research in the Alpine countries has always focused on the effects of the orography on the ambient weather and climate. In recent years, additional motivation has been provided by the following developments:
Figure 1-1. The Alps. Scale changes at
200, 500, 1000, 1500, 2000, 2500, and 3000m altitude. Terrain
above 3000m in white.
Figure 1-2. Car buried in mud after the
devastating flash flood of the river Saltina in the town of Brig,
Switzerland on 24 September 1993 (Photo: Coffrini / SonntagsZeitung
of 26 Sept. 1993)
The scientific problems pertinent to these issues are not only important in the Alpine region, but they are of general interest for many other mountainous regions of the earth, and furthermore relate to aspects of the general circulation of the atmosphere.
The Mesoscale Alpine Programme is to be undertaken as a concerted, international and interdisciplinary effort to further the basic understanding and the forecasting capabilities of the physical and dynamical processes which
These themes are intimately interrelated, and the insights gained will be mutually beneficial.
The envisioned programme includes a multi-year preparatory phase, followed by a 13-month observing period with a 3-month intensive field campaign, and will be concluded with a two-year evaluation phase.
Systematic use of operational numerical weather prediction (NWP) and research models will represent one of the key activities within MAP. The acquisition of high-resolution data sets with emphasis on moist processes and topographic circulations will help the validation of these models, and this feature is decisive for further development. These data sets will also be geared towards better understanding of the underlying physical and dynamical processes.
In addition to the operational networks, several advanced observing systems will be dedicated to the MAP Special Observing Period (SOP) with a tentative duration of 3 months. MAP will make intensive use of recent technological developments, e.g. surface based wind profilers, radar networks, airborne radar and lidar systems, light-weight sondes dropped from aircraft, and satellite products. Some of these new technologies (e.g. wind profilers) will become operationally available within the Alpine region in the next years, and will be used in conjunction with existing systems for the monitoring of longer periods (i.e. complete annual cycles) in the General Observing Period (GOP) which is tentatively scheduled to have a duration of 13 months.
The establishment of climatologies of mesoscale systems from existing archives, and the application of the most recent numerical models to data sets from previous campaigns (e.g. ALPEX and PYREX) constitutes an integral part of MAP. These activities will start well ahead of the MAP field phase, and will help the planning and fine-tuning of the observational campaign.
The Alps are very well suited for the envisioned programme, since they are very rich in topographically controlled flow features, and since they are already equipped with very dense operational networks for atmospheric observation. This includes one of the densest radiosonde networks (operated by the national weather services), as well as a precipitation network of several thousand rain gauges (operated by the region's hydrological and meteorological services).
Most national weather services in Europe utilize numerical models which include the Alpine region in their daily forecast practice. The horizontal resolution of these models ranges from ~80km to 14km (e.g. at the German and the Swiss Weather Services, see Majewski and Schrodin 1994), but some higher resolutions have also been experimentally used either in the research mode or for specific forecasting purposes. The latter includes an experimental model suite with 3.6km horizontal resolution operated by Météo France during the 1992 Olympic Winter Games in Albertville (see Bret and Bougeault 1992).
An example from the operational forecast for the flash flood case of Brig (23 September 1993; Fig. 1-3) highlights the high potential of modern numerical forecast models. Note the fairly realistic simulation of the very severe precipitation maximum near the southern border of Switzerland.
Figure 1-3. 24h accumulated precipitation
between 23 and 24September 1993, 18UTC, as predicted by the operational
Swiss Model. Isolines at 0.5, 1, 5, 10, 20, 50, 100 and 300kgm-2.
The maximum of 407kgm-2 is located over the catchment area of
the river Saltina, which resulted in catastrophic flooding of
the town of Brig, situated in the upper Rhone valley. The bold
line depicts the 750m contour of the orography. (Courtesy of Swiss
Meteorological Institute)
In general, however, snow and rainfall remain difficult quantities to forecast, especially if associated with deep convective activity and thunderstorms. Forecast failures in the Alpine region occur also as a result of the limited ability to simulate orographic circulations and their interaction with the synoptic-scale flows. For instance, while there have been significant advances in the understanding of two-dimensional flow past topography (see the contributions in Blumen 1990), there is a large gap in our knowledge of fully three-dimensional, time-dependent and moist flows over and around complex mountain ranges.
Investigations with new numerical models is expected to increase in the next few years as operational nonhydrostatic models become available. These models with horizontal resolutions of a few kilometres are currently under development at several weather services (e.g. in Canada, France, Germany, Great Britain, USA) and are also increasingly being used in a research mode (see for instance Grell and Grell 1994; Benoit et al. 1995). These new tools will enable the explicit resolution of convection, thunderstorms, and trapped orographic lee-wave features that with the present generation of forecast models are parametrized or not even represented.
With an increase in the spatial resolution of model forecasts, additional efforts will have to be undertaken to cross-validate model results and observations (Hollingsworth 1994). This calls for new developments to generate high-resolution data sets of critical atmospheric processes. It is imperative that such a data set be compiled in the context of a cost-effective international effort. Such data sets can
The flood that affected the Piedmont Region in Italy in the first days of November 1994 was one of the most damaging events that occurred in the Alpine area in this century. The heaviest rain fell in the period 4-6 November following a generally wet spell. Several stations recorded cumulated values in excess of 300mm(36h)-1 (Lionetti 1996). Two main areas and periods of precipitation can be distinguished. The first (see Fig. 1-4), characterized also by embedded convective activity (with very high peak rainfall intensities up to 50mmh-1), affected the area between the Maritime Alps / Ligurian Apennines and the Langhe hills on November 4-5. For the Tanaro river, a major right-side tributary of the river Po, the peak flow at the Montecastello gauge (7985km2) has been estimated in the range 3500 to 4800m3s-1 with a corresponding return period of 70 to 1300 years (Brath and Maione 1996). The second rainfall maximum occurred over the eastern flank of the western Alps, north of Turin, on 5-6 November. It affected the basins of several tributaries of the Po river with less severe effects. Numerical experiments have demonstrated the essential role played by the orography in determining the amount and distribution of precipitation (see for example Buzzi and Tartaglione 1995).
As a consequence of the Piedmont flood, 70 people died, and 2000 had to be evacuated. Damages to properties were extensive and estimated to exceed 20'000 billion liras (~10 billion ecu), about one third of which in public works and agriculture. For example, 150 bridges collapsed or were seriously damaged, and more than 5000 heads of livestock were lost.
The Piedmont flood event cannot be properly classified as a flash flood, since the intense rainfall period lasted for more than 24 hours and affected several medium-sized river basins. Thus it seems to be an example in which a combination of accurate meteorological short range forecasting (12-48h) by means of numerical weather prediction models, combined with multisensor (pluviometer networks, radars, satellites) real-time monitoring of precipitation and river state, and with hydrological forecasting models might have provided useful information and assistance for the public authorities. It is one primary aim of the MAP scientific community to set up and improve the basic tools needed to provide real-time flood forecasting and warning.
With respect to the mitigation of flood damages induced by heavy precipitation, a closer interaction between hydrologists and atmospheric scientists is desirable in order to define the forecast lead time and the accuracy of rainfall predictions needed to set up a useful flood forecasting, warning and response system (FFWRS) (Borga et al. 1991, Lang et al. 1996, Moore et al. 1993, Obled and Tourasse 1994). One of the most important variables influencing the adoption and success of FFWRS is the time available to issue and disseminate flood warnings and to take appropriate actions. Experience from events that occurred in the recent past suggests that, with rare exceptions, warnings cannot be really useful within a lead time of just a few hours, depending on many physical (structure of the drainage system, terrain properties, etc.) and social factors.
Figure 1-4. 1.3_1: NOAA IR (channel 5)
satellite image of 5 November 1994, 9:34UTC, taken during phase
I of Piedmont flood (Dundee receiving station).
Particular attention in the theme of flood prediction has to be paid to the link between the spatial structure of the drainage systems and that of precipitation fields. The appropriate space-time resolution for coupling the meteorological models with a rainfall-runoff scheme has to be investigated, also in terms of precipitation forecast reliability and lead time availability. The general indication is that the larger the basin area, the longer the response time. For examples, in major river basins (>10'000km2), short term flood forecasting can mainly rely upon upstream measurements of the river discharge and runoff predictions based on rainfall monitoring (Todini and Wallis 1977). In contrast, for small-scale basins (<1000km2), only accurate rain forecasts can, in principle, allow to formulate useful FFWRS. It is an open problem to what extent numerical predictions of rainfall can be improved in the near future such as to allow for sufficient time-space accuracy as needed for reliable flash-flood warnings. On intermediate scale basins (1000-10'000km2) a combination of rainfall forecast, observations and hydrological modelling is likely to give the best results.
For mid-latitude complex orography like the Alps, accurate temperature forecasts are very important as well, since required for the distinction of snowfall from rainfall. The analysis of flood events (as for instance the Brig event and the Ticino flooding of 1993, and the Reuss valley event in 1987) clearly demonstrates that the altitude at which falling snow melts during heavy precipitation periods is decisive in many basins to discriminate between flood and no-flood conditions. Additional meltwater production during warm rainy periods may also contribute to flooding situations.
The research to be conducted within MAP is relevant to global change issues in two respects.
First, mountainous regions are very effective in extracting moisture from the ambient atmospheric flow via various orographic precipitation mechanisms. Such precipitation is important not only in the considered mountainous area itself, but is often highly relevant for the fresh-water management in large neighbouring regions. In the case of the Alps, more than 100 million people rely on the Alpine rivers Rhine, Rhone and Danube for their fresh-water supply. Climatological studies, attempts to downscale global change scenarios to mountainous region, and continental-scale studies of future freshwater resources will clearly benefit from a better understanding of the pertinent meso-scale circulations and precipitation processes in mountainous regions.
Second, it has been recognized in recent years that mountains are one of the key factors in defining the geographical distribution of the climatic zones on the planetary scale. An illustrative example from a general circulation experiment is reproduced in Fig. 1-5. It shows the results of a pair of global climate simulations. In both the experiments, the geographical distribution of land and sea is prescribed, while the topography has been removed for the simulation shown in the right-hand panels. The bottom panels depict the simulated geographical distribution of wet and dry climatic zones on the northern hemisphere. The local effects of topography (such as in the Alpine precipitation anomaly) are clearly visible, but there are remarkable effects remote from mountains. For instance, large parts of Siberia are classified as dry in the topography run (in agreement with the observed climatology), but would experience substantially larger rainfall amounts in the absence of topography.
From the top panels in Fig. 1-5 it can be inferred that the control exerted by topography is through planetary-scale standing wave patterns and their embedded storm tracks and a variety of the mesoscale processes contribute to determining the precise position and amplitude of these waves. The study of these mesoscale processes is one major objective of MAP and these processes include both upper-level gravity wave drag (e.g. Palmer et al. 1986) as well as low-level drag effects (e.g. Lott 1995). These processes are subgrid-scale in current general circulation models and must be parametrized. Current schemes are not adequate, and a better understanding of three-dimensional gravity wave propagation, gravity wave breaking and flow splitting, which is anticipated by MAP, will help to improve these parametrizations.
Figure 1-5. Mean spring circulation and
precipitation patterns from global numerical simulations with
(left) and without (right) mountain topography. The top panels
show the 500hPa geopotential height (dm) and wind speed; light
and dense stippling indicates winds > 12ms-1 and > 24ms-1,
respectively. The bottom panels show precipitation rates (mmd-1);
contours are given at 1, 2, 3, 4, 5, 6, 8, 10, 15, 20 and 30mmd-1;
light stippling indicates dry regions with precipitation <
1mmd-1; dense stippling indicates wet regions with precipitation
> 3mmd-1. Taken from Broccoli and Manabe (1992).
MAP is not the first research initiative in the atmospheric sciences to focus on the Alpine region. The Alpine Experiment (ALPEX) took place 1981-82 as the last large campaign of the Global Atmospheric Research Programme with a special observing period of two months in spring 1982. The emphasis lay on general circulation aspects and on weather systems down to the meso-a scales (Gutermann and Wanner 1982; Kuettner 1986; GARP 1986). Also the modification of fronts as they approach the Alps was intensively probed during the Front Experiment 1987 (cf. Hoinka and Volkert 1992; Fronts and Orography 1992). Momentum budgets and regional wind systems over and around the simpler shaped, neighbouring range of the Pyrenees have been thoroughly investigated during PYREX in autumn 1990 (Bougeault et al. 1993). Related experiments were conducted in other mountainous regions, as for instance the Hawaiian Rainband Project (see e.g. Smith and Grubisic 1993; Rasmussen and Smolarkiewicz 1993).
Though significant advances have been made through these previous initiatives, the complexity of Alpine meteorological phenomena and the requirement for improved understanding and enhanced forecasting capabilities necessitate another concerted effort specifically addressing meso-b and smaller scale phenomena, and making use of the most recent observational and numerical tools.
It is foreseen that results from MAP will contribute to continuing research programmes. A better understanding of the dynamical effects of upper-level flow features near the tropopause will be beneficial to the interpretation of measurements obtained during the Second European Stratospheric Arctic and Mid-Latitude Experiment (SESAME), which is sponsored by the European Union. A better understanding of mountain effects in the Alpine region will also help to better assess some air-quality, transport and boundary layer problems, as for instance those addressed under the umbrella of the ALPTRAC (see EUROTRAC 1993), or POLLUMET programmes (POLLUMET 1990). The study of upper-level gravity wave breaking will touch also on aspects of interest to stratosphere-troposphere exchange studied within SPARC (1993).
All aspects regarding improved mesoscale numerical forecasts are in accord with the Short Range Weather Prediction Programme of the World Meteorological Organisation (WMO). Research tasks focusing on orographic processes influencing the regional or global climate are of high interest to the World Climate Research Programme of WMO. Therefore MAP has already been accepted as a WMO-sponsored project.
Detailed investigations concerning the generation and efficiency of convection, precipitation and hydrological processes over the Alps are of relevance to the Global Energy and Water Cycle Experiment (GEWEX). Following the presentation of MAP at the second session of the GEWEX Hydrometeorology Panel (GHP) in Toronto in August 1996, it was noted that GEWEX and especially GHP recognize the contribution of MAP activities towards improving the prediction of moist processes over and in the vicinity of complex topography, including interactions with land-surface processes. Ongoing interactions with MAP were encouraged, including cross-representation at respective working group sessions, workshops and Scientific Steering Group meetings as appropriate. This arrangement will provide for MAP to be conducted in liaison with GHP.
Finally, environmental issues are becoming more and more pressing for the Alpine eco-system as a whole (Alpine Convention 1993) to secure with the help of protocols for various fields (e.g. agriculture, traffic, tourism) this unique homeland of 11 million people and sojourn of many more visitors. Through the Mesoscale Alpine Programme, atmospheric scientists from within Europe and abroad will be able to provide more detailed background information about the weather and climate over the Alps and its impact on other components of the entire eco-system.
