Through its ability to generate large-amplitude vertical air displacements that bring moist-unstable airflow to saturation, orography serves a focus for convection initiation. As a result, orographic convection represents as an important forecasting problem and has attracted decades of intensive research. However, the physical understanding and numerical representation of orographic convection remain incomplete and elusive. This study briefly reviews the history of orographic convection research with a focus on recent findings facilitated by intensive observational field campaigns and ever-increasing computational power. It subdivides the problem into two fundamental (but not mutually exclusive) physical mechanisms: mechanically and thermally forced convection. Observational studies have investigated one or both mechanisms in increasing depth and detail over the past few decades, with distinct progress made through the Mesoscale Alpine Programme (MAP), the Convection and Orographically-induced Precipitation Study (COPS), and various others. Over the same period, the emergence of cloud-resolving simulation over large domains has permitted the explicit and systematic investigation of mountain cumuli over a spectrum of background conditions and terrain forcings. While providing critical new insights, these observational and numerical studies have also highlighted the complex interactions between airflow dynamics, cloud microphysics, and surface heat exchange that make orographic convection a highly challenging (and intriguing) problem. The study concludes with a discussion of some prominent obstacles facing the improved understanding and prediction of this problem.

Active orographic convection was observed over a small tropical island and yet a negligible amount of precipitation was observed. The generation of precipitation from a cumulus cloud depends sensitively on the nature of the updraft, the updraft aerosols, and the turbulent entrainment of dry air, all of which were observed during the DOMEX (Dominica Experiment) field campaign. Observations suggest aerosol-cloud-precipitation interactions may be important in retarding precipitation formation in thermally driven orographic convection but differences in the observed cloud layer moisture content confound the relationship. Early test simulations also suggest that the island geometry may influence the aerosol impact. This leads to the question of under what circumstances can aerosol-cloud-precipitation interactions impact orographic precipitation?

The WRF model with the aerosol-aware Thompson microphysics scheme is used to further explore the relationships between simulations with and without a surface aerosol source. Idealized cloud resolving 3D simulations with a variable sensible heat flux simulating one thermal day are performed with a set of terrain configurations, wind speeds, and cloud layer moisture contents in addition to the changing surface aerosol source. It is found that aerosols have the largest impact on precipitation when it is struggling to form. A water budget, an aerosol budget, and a close look at microphysical conversion rates from cloud to rain water help to shed light on the processes at work.

Dimensioning of flood protections is based on the estimation of the probable maximum flood (PMF). A reliable estimate of this quantity can only be made using a realistic estimate of the probable maximum precipitation (PMP) in the considered catchment. However, traditionally used procedures to estimate the PMP are not well suited for mountainous regions. These procedures typically transfer an extreme precipitation event observed in a nearby area to the catchment of interest with some adaptation for the differing topography. But the complex terrain does strongly affect the precipitation distribution and impose strong nonlinearities for the precipitation resulting from small variations in the atmospheric flow conditions. Therefore an in-depth knowledge of the precipitation characteristics of a catchment is needed to obtain realistic estimates of the PMP and eventually the PMF.

We use the high-resolution numerical weather prediction model COSMO to study small-scale processes induced by topography-flow interactions. A sensitivity analysis is performed to determine the influence of subtle variations in atmospheric parameters such as specific humidity, wind direction, and temperature on the precipitation distribution. For this purpose, various approaches are used to modify either the initial and boundary conditions of humidity and temperature. Simulations are performed for different flood events in Switzerland, including different type of synoptic forcing, such as blocked and unblocked cases, characterized by atmospheric rivers or quasi-stationary cyclone.

The results show that, for instance an increase of the specific humidity of the incident flow does not necessarily produce an increase of precipitation in the target catchment. Indeed, with increased ambient moisture, smaller mountains upstream of the catchment can be more efficient in triggering precipitation and therefore reduce the moisture available downstream. This novel approach with a set of synthetic sensitivity experiments allows estimating, for a particular catchment, the physical limits of the PMP value.

The windward slopes of the Eastern Pyrenees, as other mountainous Mediterranean regions, are regularly affected by heavy precipitation events (HPE). This study benefits from high resolution numerical simulations (with horizontal grid lengths of 2.5 km) of different case studies using the MESO-NH research model. The analysis has a twofold objective: to describe the synoptic environment in which the HPE developed and to identify the mesoscale mechanisms that lead to steady rainfall over the Eastern Pyrenees. Furthermore, the analysis has been extended to include a Lagrangian description of the flow feeding the precipitation systems which were fairly well reproduced by the model.

Results from the simulations of three different conditionally unstable episodes indicate a marked dependence of the precipitation intensity over the Pyrenees on two factors: the intensity of the wind at low and mid levels and the moisture advection towards the Pyrenees in the lowest 3 km of the atmosphere (Q3). According to the simulations, it has been detected three different rainfall intensity regimes ranging from weak to heavy orographic precipitation where Q3 exceeded 550 kg/m•s and a low level jet of 30 m/s was also present. From the backward trajectories based on Eulerian on-line tracers, it has been found that the feeding flow is confined between 0.5 and 3 km of altitude, mainly in the top edge of the conditionally unstable boundary layer (>1000 m), whereas for the precipitating systems close to the coast the flow is confined in the first 1000 m within the PBL.

An analysis is presented here of severe convection affecting the Friuli-Venezia Giulia region (FVG, northeastern Italy) during Intensive Observation Period 2b (IOP2b) in the first Special Observation Period (SOP1) of the HyMeX campaign. This work focuses on the morning of September 12, 2012, during the first of three severe convection episodes that affected the region during IOP2b. In the first episode, a supercell, which produced hail and severe damage to trees and buildings, was observed on the plain of FVG.

The same day has been analyzed already by previous studies, but here the available observations were analyzed together with new WRF model simulations, in order to identify the mechanisms responsible for the severe convection. The WRF model was used in a one-way nesting configuration, using three nested domains with grid spacing respectively of 9, 3, and 1 km.

First, the predictability of the event was analyzed. Six different simulations were performed starting at three different initial times, using respectively the ECMWF and the GFS analysis/forecasts as initial/boundary conditions. A large spread was observed among the simulations. Only a few simulations were able to reproduce intense rainfall on the plain of FVG during the morning, although with significant differences in the rainfall distribution among them. Second, it was found that in this case the GFS-initialized run starting at 12 UTC, 11 September 2012 best reproduces the observed elongation toward Slovenia of the intense rainfall maximum. The characteristics of the cell are consistent with those expected for a supercell and its simulated evolution near the Adriatic coast agrees well with the observations. Finally, the simulated evolution of some relevant instability parameters over the FVG plain and offshore (over the northern Adriatic Sea) is analyzed, finding that during this event the potential instability varies a lot even in small space and time intervals.

Rising a modest 500 m above the eastern shore of Lake Ontario, the Tug Hill Plateau of northern New York averages up to 750 cm of snow annually and likely experiences the most intense snowstorms in the world. These storms are generated by long-lake-axis-parallel (LLAP) snowbands that develop over Lake Ontario and intersect the plateau's gentle western slope. World-record snowfalls recorded on the plateau include 30 cm in 1 h (Copenhagen, NY, 2 Dec 1966) and 130 cm in 16 h (Bennetts Bridge, NY, 17–18 Jan, 1959). In addition, a remarkable 24-h accumulation of 195 cm was recorded in Montague, NY from 11–12 Jan 1997, but is unofficial because five measurements were made during the period rather than four.

During December 2013 and January 2014, the National Science Foundation sponsored Ontario Winter Lake-effect Storms (OWLeS) project examined lake-effect storms in the vicinity of Lake Ontario. Research platforms included the University of Wyoming King Air Research Aircraft, three Center for Severe Weather Research (CSWR) mobile X-band Doppler on Wheels radars, the University of Alabama at Huntsville Mobile Integrating Profiling Systems, and five mobile sounding systems. Scientists from the University of Utah and State University of New York at Albany operated a transect of profiling K-band radars from the eastern shore of Lake Ontario to upper elevations on the Tug Hill Plateau, as well as lowland (145 m MSL) and upland (385 m MSL) snow-study stations spaced ~20 km apart, providing an unprecedented look at orographic enhancement during lake-effect snowstorms.

During intensive observing periods (IOPs), snowfall rates at the upland site reached 13.9 cm/h, 36.5 cm/(6 h), and 98 cm/(24 h). During the 12-day period encompassing IOPs 1–5, the upland site received 252 cm of snow with a mean water content of only 5.9%. Despite an elevation difference of only 240 m, the mean orographic ratio (i.e., upland/lowland liquid equivalent precipitation) for 6-h periods during which manual cores of fresh snow were collected at both sites was a remarkable 2.1. Analysis of the data collected by OWLeS observing platforms combined with snow and precipitation measurements from the upland and lowland sites provides new insights into the mechanisms contributing to intense snowfall rates during lake-effect storms and a unique perspective on orographic precipitation enhancement over modest topography.

Orography downstream of large bodies of water, such as the Great Lakes of North America, often enhances lake-effect snowfall, which can be intense, extremely localized, and challenging to predict. Here we examine orographic effects during lake-effect storms over the Tug Hill Plateau, which gradually rises ~500 m above the eastern shore of Lake Ontario and receives nearly twice the annual snowfall as the surrounding lowlands. Our analysis uses data collected during the Ontario Winter Lake-effect Systems (OWLeS) field program, including S-band surveillance radar scans and data from a transect of profiling K-band radars and snow measurement stations that extended inland from the eastern shoreline of Lake Ontario to the upper Tug Hill Plateau. We specifically examine how the morphology and three-dimensional structure of lake-effect convection during OWLeS IOP2 modulates the orographic enhancement of precipitation over and around the Tug Hill Plateau.

IOP2 produced over 100 cm of snow on the western slope of the Tug Hill Plateau in 24 h. During the event significant vacillations in lake-effect morphology were observed. Organized long-lake-axis parallel (LLAP) bands centered over the transect produced the highest precipitation rates and accumulations at both lowland and upland sites, but also the lowest orographic ratios (i.e., the ratio of upland to lowland liquid precipitation equivalent). Rimed crystals and graupel were frequently observed at both sites during LLAP periods. Broad coverage periods, featuring less-organized convective coverage made up of distinct individual cells, produced lower precipitation rates at both sites but higher orographic ratios as cellular convection produced over the lake transitioned to a more stratiform mode with greater coverage and more persistent precipitation over the Tug Hill Plateau. Mixed mode periods, which consisted of broad coverage concomitant with an organized LLAP band, also exhibited broadening and a convective-to-stratiform transition. During these periods there were low orographic ratios between the upland and lowland site when the band was centered over the transect and high orographic ratios when the band was to the north or south of the transect.

Climatologically, organized bands account for 10% of lake-effect hours over Lake Ontario with broad coverage and mixed mode making up the majority of the remainder. Accordingly, this study suggests that the Tug Hill Plateau precipitation maximum may be largely due to the broadening of precipitation coverage and frequency over the uplands rather than the invigoration of LLAP bands over the Tug Hill Plateau.

The Tug Hill Plateau of upstate New York rises ~500m in elevation east of Lake Ontario, observes frequent lake-effect snowfall, and is one of the snowiest locations in the eastern United States. This, in addition to its simple geometry and the weather surveillance radar situated atop it, make the plateau a favorable location to study orographic precipitation. Using a variety of datasets including imagery from the KTYX WSR-88D radar, this work examines the climatological characteristics and orographic enhancement of lake-effect precipitation over the Tug Hill Plateau over 13 cool seasons (16 September 2001 - 15 May 2014).

Lake effect is important to the hydroclimate of the region, with lake-effect days accounting for 61-76% (24-37%) of the mean cool-season snowfall (liquid precipitation equivalent). The greatest lake-effect snowfall occurs over the western and upper plateau, decreasing abruptly in the lee over the Black River Valley.

Wind speed and direction are shown to strongly affect the precipitation distribution over the plateau. Low wind speeds produce a maximum in radar echoes over the lower slopes of the plateau, and increasing speed shifts the maximum to the higher slopes the plateau. During the highest winds, frequent precipitation persists well inland into the western Adirondack Mountains. A pronounced maximum in echo frequency is present over the plateau under westerly winds (260°-280°), but even when more northerly winds (281°-300°) displace the primary lake-effect structures south of the plateau, precipitation nevertheless continues to initiate or intensify over the plateau.

Orographic enhancement is also examined in the context of the morphological characteristics of lake-effect precipitation features. The two most common morphological types, broad coverage and LLAP, account for 72% and 24% of the total time of lake-effect periods, respectively. Broad coverage exhibits greater enhancement over the plateau than LLAP, despite the fact that LLAP bands produce the most intense snowfall. Collectively, these results illustrate the unique characteristics of orographic precipitation enhancement over the relatively modest topography of the study region during lake-effect storms.

Gravity waves forced by terrain-induced vertical displacements of stably-stratified air parcels are referred to as mountain waves. There has been considerable investigation of the role of mountain waves in downslope windstorms, clear-air turbulence and orographic drag influence on the general circulation. In contrast, the influence of mountain waves on orographic precipitation has received comparatively little attention in the literature. Most of these investigations have employed a simulation approach using either quasi-analytical linear models or numerical weather prediction models. Observational studies linking mountain waves to orographic precipitation are scarce. Doppler-radar-based studies of mountain waves and orographic precipitation have provided useful insights, but are limited by not describing the field of vertical air motions across the primary barrier through the full depth of precipitating cloud. It is these vertical motions acting in concert with horizontal motions that directly impact the spatial distribution of precipitation relative to the barrier.

The present study addresses these limitations through analysis of a winter storm passing over the Park Range of northern Colorado on 15 December 2010 during the Colorado Airborne Multi-Phase Cloud Study (CAMPS) field experiment. Observations from the W-band Wyoming Cloud Radar onboard the University of Wyoming King Air (UWKA) research aircraft are used to document horizontal and vertical motions along with precipitation structure in a two-dimensional vertical plane that extends across the Park Range near Steamboat Springs from upstream of the windward slope, over the crest and downstream of the lee slope. Additional observations employed in the analysis include those from the UWKA flight-level instrumentation and ground-based observations from the Atmospheric Radiation Measurement (ARM) program mobile facility (AMF2) such as a W-band scanning radar, balloon sounding system and surface meteorology instrumentation. Surface observations from the Desert Research Institute (DRI) Storm Peak Laboratory (SPL) are also incorporated into the analysis.

The winter storm investigated in this study is associated with a general zonal flow over the western continental U.S. and significant vertical wind shear between 700 and 500 hPa. A vertically-propagating wave forced by the Park Range is most evident above 4 km MSL and associated with relatively-wide, upstream-tilted updrafts and downdrafts located above the Park Range windward and lee slopes, respectively. The Park Range also forces a trapped-lee wave that manifests itself as a relatively erect updraft ~15-20 km east of the crest. Smaller-scale trapped lee waves forced by terrain upstream of the Park Range are evident below 4 km MSL and associated with rotor circulations composed of relatively-narrow, updrafts and downdrafts located above the Yampa Valley and the Park Range windward slope. A ~1 km-thick layer of strong vertical shear exists between the mountain waves forced by the Park Range and those forced by upstream terrain. This shear layer exhibits a large vertical displacement over the Park Range, with relatively-strong westerly winds plunging to low levels over the lee slope. While precipitation on the Park Range windward slope is generally enhanced for the event, data analysed for this case surprisingly does not show a spatially and temporally consistent correlation between mountain-wave kinematic structures and orographic precipitation. Transient processes such as frontally-forced mesoscale circulations, precipitation bands, and/or wave-regime interactions may have masked this correlation.

Major precipitation events occur in the Banff/Calgary area of Alberta that are a critical aspect of the region's water cycle and they can lead to major disasters such as the June 2013 flooding. This extreme event was the most costly natural disaster in Canadian history. Such events can bring rain, snow or both to the area but there has never been an atmospheric-oriented special observation of these events beyond those made with operational networks. The objectives of this field experiment are to better document and understand precipitation and associated atmospheric conditions in the Banff-Calgary region. To study such storm, based on the climatology of the area, intensive observational period were conducted in March and April 2015. These included detailed measurements of precipitation and weather conditions at the surface, vertical Doppler radar observations, soundings, and profiles down a mountainside examined during each weather event. An overview of the weather events and key preliminary findings will be presented using both observations and model information with a special focus on the characteristics of the transition between rain and snow which often occurs in this area in the spring.

The continuing increasing size of weather radar archives offers an unique opportunity to investigate the statistical properties of precipitation.

In this study, quality-checked quantitative precipitation estimates produced by MeteoSwiss by combining radar and rain-gauge measurements for the period 2005-2014 are employed to investigate the influence of different spatial and temporal scales on the statistics of heavy precipitation in Switzerland.

Intensity-Duration-Frequency (IDF) curves for precipitation extremes measured over several spatial and temporal scales are shown, and compared with those derived from point rain-gauge measurements. The spatial continuity of regional rainfall distributions and IDF curves in Switzerland is also analyzed, thus permitting to identify regional differences in the behavior of heavy rainfall.

This study constitutes the scientific framework necessary to identify optimal thresholds of precipitation accumulations for a nowcasting system recently developed at MeteoSwiss and specifically designed to issue heavy rainfall alerts over pre-defined geographical regions.

Ice nucleation and cloud droplet freezing are two of the most poorly constrained cloud microphysical processes in the atmosphere. Orographic wave clouds at mid-tropospheric levels provide an ideal testbed to study these processes by combining aircraft measurements and numerical modeling. In the presented work we investigate wave clouds observed over the Rocky Mountains during the ICE-L campaign using the Unified Model. The thermodynamic and dynamic properties of the wave clouds are extremely well captured in the simulations, which allows us to focus on the representation of cloud and aerosol processes. Cloud microphysics are represented with the newly developed CASIM microphysical scheme, which accounts for aerosol processing and vertical transport. We evaluate the performance of different ice nucleation parameterizations in representing the measured microphysical evolution. In addition the vertical redistribution of aerosol particles due to sedimenting ice crystals is investigated. This comparison provides important insights in our current understanding of cloud droplet freezing and ice nucleation and the role of mid-tropospheric clouds to redistribute aerosol particles.

A high-resolution wireless network of automatic meteorological weather stations (Vaisalla WXT520, Finland), about 200m spaced, was installed in a 12 km² watershed, in a hilly landscape, with altitudes varying between 1100 to 1400 m, at Extrema city, MG state, in Brazil, with the purpose to measure the spatio-temporal variability and mesoscale processes, which includes the heterogeneities of air temperature and humidity, and wind circulations controlled by the topography. The catchment is strategically important, located at the headwaters of the Cantareira reservoir system, which provides half the water for the São Paulo megacity. The measurements will help to improve specification of initial conditions in Regional Climate Models designed for weather forecasting, building of hydroclimatological data set used in models of catchment water balance, and investigation of the effect of global climate changes in mountain regions, among others. First measurements suggest the occurrence of 3m/s up-slope wind (1 m/s down-slope), with regimes of temperature 2°C warmer (4°C colder) on the valley during the daytime (night-time). The air humidity is higher down the valley and reaches two maxima at 12am and 7pm, and minima at 6am and 3pm. Despite it is a small watershed, previous data collected during four years with 4 rain gauges show a difference of annual rainfall of about 200 mm/yr higher in the headwaters compared to the mouth of the stream. The basin is North-South orientated, and is affected by differential heating between the West and East slopes of about 1.5°C. We show spatio-temporal patterns of weather variables for different synoptic conditions, specially in pre-cold front, post-cold front systems and along the ZCAS rainy events (South Atlantic Convergence Zone). We run a very high resolution (100m) Advanced Regional Prediction System (ARPS) model, using the Large Eddy Simulation (LES) mode, and compared to observations. The calculations represented well the local wind circulation and the patterns of temperature and humidity within the watershed.

Banner clouds are clouds in the lee of steep mountains or sharp ridges. Previous work suggests that the main formation mechanism is vertical uplift in the lee of the mountain. On the other hand, little is known about the Lagrangian behavior of air parcels as they pass the mountain, which motivates the current investigation. Three different diagnostics are applied in the framework of Large Eddy Simulations of air flow past an isolated pyramid-shaped obstacle: Eulerian tracers indicating the initial positions of the parcels, streamlines along the time-averaged wind field, and online trajectories computed from the instantaneous wind field.

All three methods diagnose a plume of large vertical uplift in the immediate lee of the mountain. According to the Eulerian tracers the cloudy parcels originated within a fairly small coherent area at the inflow boundary. In contrast, the time-mean streamlines indicate a bifurcation into two distinct classes of air parcels with very different characteristics. The parcels in the first class originate at intermediate altitudes, pass the obstacle close to its summit, and proceed directly into the cloud. By contrast, the parcels in the second class start at low altitude and take a fairly long time before they reach the cloud on a spiralling path. A humidity tracer quantifies mixing, revealing partial moistening for the first class of parcels and drying for the second class of parcels. For the online trajectories, the originating location of parcels is more scattered, but the results are still consistent with the basic features revealed by the other two diagnostics.

Banner clouds are clouds that appear on the leeward side of steep mountains and sharp ridges on otherwise cloud-free days. Previous work has shown that the main formation mechanism is strong lifting of air and associated adiabatic cooling on the leeward side of the mountain.

Banner clouds can frequently be observed at mountains like Matterhorn in the Swiss Alps or Mt. Zugspitze in the Bavarian Alps. A few years ago systematic observations were carried out at Mt. Zugspitze. It turned out that banner clouds at this moutain have a preference for south-easterly flow and appear to avoid northerly flow. In addition, numerical simulations with idealized mountains indicated that weak stratification as well as a rather steep mountain are favourable for the formation of banner clouds.

Here we present numerical simulations in order to understand why banner clouds at Mount Zugspitze have a preference for south-easterly flow. To this end we carried out large eddy simulations with a special focus on the shape of the orography. In a first set of simulations we used idealized orography while systematically varying the windward-leeward asymmetry of the obstacle. Banner clouds we readily formed for mountains with a steep slope on the leeward side and a flat slope on the windward side, because in this setup a bow-shaped vortex can form on the leeward side; the latter is important as it produces the vertical uplift and associated adiabatic cooling. In a second set of simulations we used the realistic orography of Mt. Zugspitze and its environment while varying the wind direction of the incoming flow. In these simulations we found that banner clouds do preferably form for south-easterly flow while they tend avoid northerly flows. This result is in qualitative agreement with both the observations and our idealized simulations. It is concluded that the preference of banner clouds for south-easterly flow at mount Zugspitze must be due to the asymmetry of the orography.

Radiation fog is a high impact weather type which remains extremely difficult to forecast accurately, despite recent advances in numerical weather prediction. Although the basic requirements for fog formation of high relative humidity, low wind speed, and clear night skies are well known, the more complex local and non-local processes which influence the timing and location of fog formation are not well understood.

In order to better understand these processes, and to improve the fog-forecasting ability of the Met Office Unified Model, a large observational campaign (LANFEX) is currently being undertaken at two locations in the UK, running over two winters. The campaign comprises 13 sites at a variety of hilltop and valley sites in Shropshire, as well as five companion sites in the much flatter Bedfordshire area in eastern England. The six main sites are equipped with comprehensive instrumentation including flux towers, nephelometers, aerosol spectrometers, and dewmeters. Further to this, 15 sites are also equipped with newly developed fog spectrometers for measuring droplet size distributions and the liquid water content of fog.

The sites in Shropshire are located within a small area (largest distance between sites is 27 km) and were chosen to sample contrasting terrain types – particularly with regard to valley geometry and site elevation. Initial results from the campaign will be presented, focusing in particular on the observed differences in fog formation and development between the various sites.

Thermally driven flows over simple slopes are a relevant research topic, not only per se, but also as a source of key concepts for understanding and modelling many other flows over more complex topographies. However, compared to down-slope, up-slope flows have received much less attention in the literature. Indeed, to investigate katabatic winds many extensive and well equipped field measurements were performed in recent years under various research projects, and a series of high-resolution numerical simulations were run. On the contrary, few field experiments have provided detailed datasets documenting the development of anabatic flows, and the analysis of numerical investigations still relies on Schumann’s (1990) pioneering LES simulations. Also, analytic solutions - such as Prandtl’s (1942) constant-K profiles - reproduce fairly well katabatic flows, but are definitely inadequate to accurately reproduce field data for up-slope flows (Defant 1949).

In particular, some open questions still claim for further investigations, such as the conditions of instability of slope-parallel flow vs. vertical motions, and the related possible occurrence of flow separation, and the similarity analysis of slope-normal velocity profiles of temperature anomaly, wind intensity and turbulence related quantities.

Here a review of the state of the art on the subject is proposed, along with some insights into possible future developments.

References

Defant, F., 1949: Zur Theorie der Hangwinde, nebst Bemerkungen zur Theorie der Berg- und Talwinde. [A theory of slope winds, along with remarks on the theory of mountain winds and valley winds]. Arch. Meteor. Geophys. Bioclimatol., Ser. A, 1, 421-450 (Theoretical and Applied Climatology). [English translation: Whiteman, C.D., and E. Dreiseitl, 1984: Alpine meteorology: Translations of classic contributions by A. Wagner, E. Ekhart and F. Defant. PNL-5141 / ASCOT-84-3. Pacific Northwest Laboratory, Richland, Washington, 121 pp].

Prandtl, L., 1942: Strömungslehre [Flow Studies]. Vieweg und Sohn, Braunschweig, 382 pp.

Schumann, U., 1990: Large-eddy simulation of the up-slope boundary layer. Quart. J. Roy. Meteor. Soc., 116, 637-670.

The breakup of a nocturnal temperature inversion during daytime is studied in an idealized valley by means of high-resolution numerical simulations. Vertical fluxes of heat and mass are strongly reduced as long as an inversion is present, hence it is important to understand the mechanisms leading to its removal. In this study breakup times are determined as a function of the radiative forcing. Further, the effect of the nocturnal inversion on the vertical exchange of heat and mass is quantified. The Weather Research and Forecasting (WRF) model is applied to an idealized quasi-two-dimensional valley. The net short-wave radiation is specified by a sine-function with a amplitude between 150 and 850 W m−2 during daytime and is zero during the night.

The valley inversion is eroded within five hours for the strongest forcing. A minimal amplitude of 450 W m−2 is required to reach the breakup, in which case the inversion is removed after eleven hours. Depending on the forcing amplitude, between 10 and 57% of the energy provided by the surface sensible heat flux is exported out of the valley during the whole day. The ratio of exported to provided energy is approximately 1.6 times larger after the inversion is removed than before. More than five times the valley air mass is turned over in twelve hours for the strongest forcing, while the mass is turned over only 1.3 times for 400 W m−2 . An even smaller forcing amplitude leads practically to a decoupling of the valley air from the free atmosphere.

High-resolution numerical simulations performed with the Weather Research and Forecasting (WRF) model are analyzed to investigate the atmospheric boundary layer (ABL) structures associated with the development of a lake-breeze and valley-wind coupled system developing in the southeastern Italian Alps, the so-called “Ora del Garda” wind. Three mesoscale domains, forced by reanalysis data field, are used to drive the finest domains, in which the large-eddy technique is used, achieving a final horizontal resolution of 80 m over three different target areas. Model results complement an existing dataset composed of a series of measurement flights and surface observations. The flights explored specific valley sections at key locations in the study area, namely over the lake’s shore, at half valley and at the end of the valley where the breeze blows. Model results display a good agreement with the experimental dataset. In particular, the surface diurnal cycles of radiation, wind and air temperature are satisfactorily reproduced. The typical structure of the valley ABL, characterized by shallow or even absent mixed layers surmounted by slightly stable layers extending up to the lateral crest level, due to compensating subsidence in the valley core, is also well reproduced in the simulated fields. Moreover, the simulations confirm characteristic local-scale features of the thermally-driven wind field suggested by the analysis of the airborne dataset as well as from previous observations in the area. For example, the model shows the development of inhomogeneities in the cross-valley thermal field, caused by the propagation of the lake breeze and by the different heating between the sidewalls of the valley, as well as the formation of a structure resembling a hydraulic jump in the area where the Ora del Garda flows down into an adjacent valley from an elevated saddle.

In urbanised alpine valleys, persistent stable conditions, as they may occur in winter, lead to high pollution levels. Under these conditions, the atmospheric circulation within the valley is most often decoupled from the synoptic winds and the circulation in the valley is controlled by thermal (along-slope and along-valley) winds. In the simplest case of an idealised valley with uniform surface properties, the intensity and direction of these winds are controlled by the variations in the stratification within the valley, which ultimately result from the variations of the underlying terrain. These observations motivate the present study, which aims at investigating the atmospheric circulation and the heat budget of an idealised valley that opens either on a narrower valley (pooling case) or on a larger valley (draining case). The opening of the idealised valley on a plain is used as a reference case. The WRF numerical model is used to quantify the differences in the nocturnal heat budget between these different configurations.

The numerical results show that the down-valley winds are weaker for both the draining and pooling cases than for the reference case. The reason for this reduction in the intensity of the down-valley winds is that the hydrostatic pressure difference close to the surface between two valleys is weaker than between a valley and a plain.

Because the pressure gradient is related to the difference in the thermal structure of the valley boundary layers, we next consider the heat budget equation averaged over the valley and compare the terms of this equation in each valley for the pooling and draining cases.

Interestingly, after a transient stage associated with the growth of the down-valley winds, a quasi steady state is reached in all valleys, with almost the same valley-averaged instantaneous cooling rate.

However, the details of the transient and steady states are different between the pooling and draining cases. For the draining case, the cooling rate is similar to that of the reference case for both the transient and steady states, with only minor differences in the advection contribution due to the weaker down-valley winds. In contrast, for the pooling case the cooling rate during the transient stage is 60% larger than for the reference case because of large differences in the advection contribution. When the steady state is reached, the dynamical contribution (i.e. the sum of the advection and the sensible heat flux contributions) to cooling is almost identical.

After the down-valley winds have developed, we observe that the cooling rate hardly varies along the valley axis for both the pooling and draining cases. This finding is discussed using volume arguments.

Mountains and coastlines both favor the development of thermally-driven diurnal circulations, including slope flows induced by elevated heating and land/sea breezes generated by differential heating of land and water bodies. Although slope flows and sea-breeze circulations have each been studied intensively using observations and numerical simulations, understanding of the interactions between the two remains limited. In particular, some studies suggest an enhancement of the sea breeze in the presence of coastal orography while others point to a blocking or weakening effect. To gain insight into these interactions, we conduct a series of large-eddy simulations of daytime airflow over an idealized Gaussian-shaped island terrain. We analyze the simulated sea-breeze propagation characteristics and frontal-circulation strength under varying environmental and topographic factors, including island geometry (width, terrain height and steepness), boundary-layer static stability, and ambient winds. The results suggest that inland orography accelerates the sea-breeze front inland but also causes a weakening of the baroclinicity and frontal circulation. Over sufficiently tall mountains, the latter effect causes the sea-breeze front to vanish entirely as it ascends the slope. The mountain effects on the sea breeze are quantified by tracking the frontogenesis terms and along-front vertical motion in a sea-breeze-following reference frame. This analysis suggests that the mountain upslope flow acts much like an onshore ambient wind: it hastens the inland frontal propagation but also induces strong frontolysis, both of which scale with the mountain height. The potential for sea-breeze blocking at the foot of the mountain is also assessed through a simple Boussinesq scaling.

The regional scale thermally driven circulation between the European Alps and the alpine fore-land - named Alpine pumping – occurs regularly under clear and calm weather conditions. While previous studies focused on Alpine pumping impacts on moist convection and transport of air pollutants over the mountains, this study was motivated by its contribution to the ventilation of the city of Munich, located about 50km north of the European Alps in undulating and only slightly inclined terrain where local thermal circulations are weak. Hourly and daily data from a reanalysis driven regional climate simulation using the model COSMO-CLM are used to identify days with Alpine pumping and to determine the mean diurnal variation of the direction, intensity, and extension of the regional thermal circulation. The model domain extends from the foothills of the Swabian Mountains and the Bavarian Forest in the north to the southern side of the Alpine divide. The model simulation for the years 1989 to 2008 consists of 110x112 horizontal grid points with 2.8km resolution and 50 non-equidistant vertical levels with higher resolution near the surface.

Four literature based combinations of meteorological criteria have been tested to identify days favourable for Alpine pumping from COSMO-CLM results. The first criterion selects days with a daily sum of solar radiation greater or equal 20MJ/m2 and has also been used in an earlier observational study. The mean annual number of 60 days fulfilling the criterion in the model simulation compares well to the 67 days per year determined from observations. The other three criteria combinations consider a maximum upper level wind velocity, a maximum daily precipitation sum, and/or a maximum mean cloud cover. The mean annual number of selected days is lower for these criteria combinations and ranges between 20 and 52.

The daily solar radiation sum of 20MJ/m2 is only exceeded during the months from April to September with highest exceedance frequencies from May to August. In contrast the criteria combinations without a radiation threshold occur all year round. In agreement with observations, the simulated regional thermally driven wind field extends up to about 100km north of the Alps, with average near surface wind speeds of around 1 to 1.5 m/s in the Munich area. With increasing distance from the Alps the diurnal cycle of alpine pumping is delayed by up to 3h. The simulated mean depth of the daytime inflow layer derived from meridional vertical cross sections of the wind field through Munich ranges between 500m and 1500m, while the depth of the nocturnal outflow layer typically reaches up to a few hundred meters.

The DEEP propagating gravity WAVE program (DEEPWAVE) is a comprehensive, airborne and ground-based measurement and modeling program centered on New Zealand and focused on providing a new understanding of gravity wave dynamics and impacts from the troposphere through the mesosphere and lower thermosphere. This program employed the NSF/NCAR GV (NGV) research aircraft from a base in New Zealand in a 6-week field measurement campaign in June-July 2014. The region near New Zealand was chosen since all the relevant gravity wave sources (e.g., mountains, cyclones, jet streams) occur strongly here, and upper-level winds in austral winter permit gravity waves to propagate to very high altitudes. During the field phase, the NGV was equipped with new Rayleigh and sodium resonance lidars and a mesospheric temperature mapper, a microwave temperature profiler, as well as dropwindsondes and a full suite of flight level instruments providing measurements spanning altitudes from immediately above the NGV flight altitude (~13 km) to ~100 km. The DLR Falcon, equipped with a down-looking Doppler wind lidar and in situ instrumentation, was also available for coordinated flights. The ground based instrumentation included a wind profiler and radiosonde unit on the west coast of the South Island, as well as ground based lidars and high-altitude radiosonde sites.

Highlights of observations of deep propagating gravity waves during the field phase will be presented. During DEEPWAVE, 16 Intensive Observation Periods (IOPs) occurred, which featured 26 NGV and 13 Falcon research flight missions, as well as a comprehensive suite of ground based observations. These observations include cases of topographically generated gravity waves, as well as waves generated by generated by non-orographic sources such as jet streams and cyclones. Examples of coupling between the lower and upper levels will be shown using the in situ and remote sensing observations, as well high-resolution models with deep domains.

Results will be presented from high-resolution and adjoint models with deep domains for cases over New Zealand during DEEPWAVE that feature GW wave propagation. Comparisons will be made with DEEPWAVE observations. These high-resolution models highlight the role of lateral shear from the jet stream that refracts vertically propagating gravity waves generated by regions of high terrain, such as New Zealand. The predictability links between the tropospheric fronts, cyclones, jet regions, and GWs that vertically propagate upward through the stratosphere are quantified using a nonhydrostatic adjoint model. Preliminary results will be presented that quantify the degree to which forecasts of GW launching and characteristics are sensitive to the model initial state and in particular to synoptic-scale and mesoscale characteristics of mid-latitude cyclones.

A simple conceptual model of atmospheric rotors in the lee of a mountain range emerged from the first organized glider observations as early as the 1950s, during the Sierra Wave Project, and was refined by in situ research aircraft measurements over the Colorado Rockies in the 1970s. Since then, rotors have been traditionally depicted as turbulent horizontal vortices in the lee of and parallel to mountain ridges, forming in association with large-amplitude mountain waves.

The Terrain-induced Rotor Experiment (T-REX, Owens Valley, California, 2006) was the most recent coordinated effort to study the coupled system of mountain waves, boundary layer and rotors. The comprehensive T-REX observational dataset, consisting of ground-based and airborne in situ and remote sensing measurements, provides the opportunity to reconsider the rotor concept in a deep valley.

In this context, the idealized picture of a rotor proves to be simplistic and needs to be extended taking the characteristics of the valley environment into account. For example, rotor formation may be influenced by factors like the onset of thermally-driven slope and valley flows, terrain-forced flow channelling, the presence of a secondary parallel ridge, or the along-ridge variation of crest height.

A rigorous re-analysis of T-REX cases with a strong flow response to the Sierra (IOPs 1, 3, 4, 6, and 13) has been carried out. The analysis reveals that, at any given time, multiple processes may contribute to the generation of severe turbulence in the valley. For example, at nighttime, both wave-induced pressure gradients and positive buoyancy forces may work in concert to make the flow separate well above the valley floor, leading to enhanced turbulence in the valley that bears similarities in structure and strength to the classical depiction of a rotor but does not extend to bottom of the valley. Beyond rotors, turbulent interaction of cross-mountain wind, intruding into the valley, with strong pressure-driven channelled up-valley flow can give rise to turbulence intensities as large as in rotors. Based on evidence from these and other cases, we propose extensions to the current conceptual model of an atmospheric rotor that are appropriate for a deep valley.

Lee rotors are elongated vortices formed beneath trapped lee waves, where the flow stagnates and reverses near the surface, often originating flow unsteadiness and turbulence. These flow structures can constitute a serious aviation hazard, however, they are hard to forecast, because their scale is relatively small (a few km across), therefore being inadequately resolved by both global and even some regional weather prediction models. The nature of rotors, with flow stagnation and reversal, suggests that they are intrinsically nonlinear phenomena. Their accurate forecast is probably only possible using high-resolution numerical simulations, with grid spacings smaller than 1km, and resolving part of the turbulence (e.g. LES), because small-scale structures, and their impact on the large-scale flow, are intrinsically 3D, even when the large-scale flow, and the orography that generates it, are 2D. However, the fact that the near-surface flow stagnation leading to rotors has been observed for relatively modest mountain heights suggests that the flow may be nearly linear outside the boundary layer in many situations of practical interest, and only boundary layer effects induce nonlinearity, with frictional effects facilitating flow stagnation. On the other hand, it was seen previously that linear theory can give valuable qualitative indications about the onset of flow stagnation outside the boundary layer. This suggests the approach used in the present study, where the linear model of trapped lee waves over 2D obstacles developed by Teixeira et al. (2013) is extended to include a very simple representation of the boundary layer, following Smith et al. (2006). The drag calculations of Teixeira et al. (2013) found a significant correlation between trapped lee wave drag and the onset of rotors. Here we will go further by calculating the conditions under which flow stagnation occurs in the lee of 2D ridges. The flow stagnation condition is diagnosed based on the streamwise velocity perturbation (which is closely related to the drag). The existence of a boundary layer promotes flow stagnation by concurrently amplifying this velocity perturbation and attenuating the mean incoming flow. The predictions from this linear model are compared with the regime diagrams of Vosper (2004) and Sheridan and Vosper (2006), showing broad consistency.

The formation of rotors and their coupling to waves aloft is investigated using a numerical model and linear wave theory. The study considers a set of uniformly stratified flows, defined by a range of stabilities and mountain widths, over an isolated mountain ridge. The resulting wave regimes range from weakly to strongly non-linear and hydrostatic to non-hydrostatic. Working in conjunction with surface friction, mountain waves give rise to boundary layer separation and rotors in a broad range of simulated flows and conditions.

The deepest rotors and strongest reversed flows at the ground are found to develop in non-linear and moderately non-hydrostatic cases. This finding agrees with linear theory, which predicts mountain wave amplitudes to be largest in such flows. However, in near-hydrostatic conditions, for which linear estimates of wave perturbations are comparatively small, boundary-layer rotors also form. These rotors do not directly adjust to the flow structure aloft and are decoupled from it by wave breaking

It is shown, by means of linear interfacial wave theory, that an undular bore that forms on a strongly-stratified low-level jet can explain the presence of rotors in hydrostatic flows. Part of the wave energy trapped at the interface can still leak through the evanescent layer that owes its origin to wave breaking. In the process, short-wavelength vertically-propagating modes are generated, which distinctively impact the flow structure of the stratified atmosphere aloft.

Our results suggest that rotors in uniformly stratified flows owe their origin to non-hydrostatic wave modes that are either part of the primary wave field or represent secondary wave motions that are generated by breaking of near-hydrostatic waves.

Numerous studies have shown inversions to be importan for the creation of perturbations in atmospheric flows in the vicinity of mountains. Here, several aspects of the importance of inversions are presented. They are based on both observations and numerical simulations of features ranging from gravity waves with extreme rotors to weak mountain flow. A preliminary attempt to map the climatology of inversions in Iceland is also presented. The above studies indicate that elements of the flow may be very sensitive to not only the elevation, but also the strength or sharpness of the inversion. A very little temperature change at the inversion may transform the downstream flow pattern, leading to an increased magnitude of vertical velocities by an order of magnitude. Numerical weather prediction models tend to fail in reproduction of the inversions.

A combination of real and virtual topography, as opposed to the real or actual topography alone, describes the essentials of stratified flow over mountain ranges and leeside valleys or plains when a layer capped by a strong density step exists above the topography. This cap acts as virtual topography for the stratified flow aloft and will control its response.

On 14 March 2006 [Intensive Observation Period 4 of the Terrain-Induced Rotor Experiment (T-REX)], a nearly neutral cloud-filled layer, capped by a strong density step overflowed the Sierra Nevada and separated from the lee slope upon encountering a cooler valley air mass. The flow in this lowest layer was asymmetric across and hydraulically controlled at the crest. The density step at the top of this flowing layer formed a virtual topography, which descended 1.9 km and determined the horizontal scale and shape of the flow response aloft reaching into the stratosphere. A comparison shows that the famous, 11 January 1972, Boulder windstorm case was similar: hydraulically controlled at the crest with the same strength and descent of the virtual topography. In both of these cases the layer in contact with the real topography makes a transition from subcritical to critical flow at the crest and becomes supercritical downstream. It becomes subcritical again farther downstream after transitioning through a hydraulic jump. For the less-well-known Boulder case of 18 February 1970 the layer in contact with the real topography was subcritical everywhere with only a slight dip above the crest, typical of subcritical flows. In all three cases the shape of the virtual topography was very different than the real topography.

The response aloft in the troposphere to the virtual topography differed among these three cases. This response can be analyzed with the internal Froude numbers of the troposphere layer and the virtual topography beneath these layers. For the Sierra case the troposphere layer is subcritical and hence decelerates and slows with a decrease in the height of the virtual topography. This is a subcritical response with an associated rise in the elevation of the tropopause. Although the famous 11 January 1972 Boulder case has virtual topography that is similar to that of the Sierra case, the troposphere layer makes a transition from subcritical to supercritical flow. The difference is in the existence of a weak cap aloft bounded by a stagnant isolating layer. The 18 February 1970 Boulder case had no downslope windstorm and no response in the supercritical flow in the troposphere layer aloft, despite the 80 m/s westerly wind component at 10 km.

Situated on an artificial island surrounded by complex mountainous orography, the Hong Kong International Airport (HKIA) is susceptible to the occurrence of terrain-induced airflow disturbances under suitable meteorological conditions. These airflow disturbances are known to bring about windshear and/or turbulence when encountered by landing/departing aircraft, and pose potential aviation safety hazard.

During a particular episode in March 2015, the arrival of an intense late-season northeast monsoon resulted in over 60 reports of windshear encounter at HKIA within a day, necessitating a number of aborted landings and even diversion in some cases. Analysis of wind field around HKIA, as visualised by Doppler LIDAR PPI scans, revealed typical flow patterns conducive to terrain-induced windshear for sustained periods of time, including occurrence of a low-level southeasterly jet aloft the arrival corridor and subsequently frequent shedding of vortices downstream of the Lantau mountains as winds veered south.

This study examines the performance of the Hong Kong Observatory (HKO)’s Aviation Model (AVM) in capturing such terrain-induced windshear features through comparison with aircraft pilot reports, aircraft Quick Access Recorder (QAR) data and LIDAR observations. The AVM is a sub-kilometre implementation of the Weather Research and Forecast (WRF) model by HKO in an operational setting, providing hourly-updated short-term forecasts around HKIA at horizontal resolution down to 200 m. Simulated LIDAR return by the AVM, both in the form of PPI scans and aircraft glide-path headwind profiles, will be studied with a view to identifying the strengths and weaknesses in reproducing the spatio-temporal characteristics of the observed wind features. Sensitivity of the simulated wind features to horizontal resolution of the model domain, as well as predictability of such features, will also be briefly discussed.

Numerical weather prediction (NWP) models are providing weather forecasts at increasingly higher resolutions, allowing boundary layer phenomena to be resolved over areas with mesoscale topography. Multi-year climatographies, constructed from high-resolution NWP model output, can therefore now provide an improved representation of the spatiotemporal structure and other characteristics of these phenomena, including atmospheric boundary layer (ABL) heights. Knowledge of the ABL height variability is important for many studies including the dispersion of air pollution, weather forecasting, and planning of boundary layer field experiments. In this paper, two years of hourly output from an operational forecasting model run with a 1.1-km grid interval are analyzed to construct a daytime ABL height climatography and to investigate the spatiotemporal behavior of ABL heights over an area of complex terrain in northwestern Utah, centered around an isolated mountain. The two year climatography shows considerable variability in ABL heights, which generally increase from west to east and from north to south. The variability cannot be solely explained by gradients in sensible heat flux due to heterogeneities in surface type. Changes in terrain elevation and associated thermally driven circulations are partially responsible for the ABL height variability. Data from intensive observational periods during the MATERHORN field campaign are used to demonstrate the complexities associated with comparing simulated PBL heights to observations. This study demonstrates that climatographies constructed from high-resolution NWP output can potentially provide useful information about ABL height variability for different applications, and can improve our understanding of boundary layer structure over complex terrain.

Urbanised Alpine valleys experience high pollution levels during wintertime stable conditions. One notable example is the Arve valley around the small city of Passy, located in the French Alps 20 km west of Chamonix, which is one of the most polluted places in Région Rhône-Alpes under such conditions. A key controlling factor of pollution lies in cold air pools that form at the bottom of the valleys. Only recently have field campaigns been specifically focussing on such conditions, e.g. COLPEX (Cold-air Pooling EXperiments) in the Clun valley in England, I-Box (Innsbruck Box) in the Inn valley in Austria and PCAPs (Persistent Cold-Air Pool Study) in the Salt Lake basin in the USA.

A scientific project devoted to the study of the atmospheric circulation in the Passy valley has been under way since January 2014, named the Passy project, in close connection with a programme devoted to the study of PM10 distribution in the same valley (managed by the air quality agency of Région Rhône-Alpes and laboratory LGGE in Grenoble). The Passy project has two main objectives: (I) to characterise the atmospheric circulation and the thermal structure of the inversion layer within the valley during stable conditions; (ii) to quantify their impact on pollutant dispersion. To start with, a field campaign was held in winter 2014-2015, during which extensive meteorological data and supplementary atmospheric composition data (mostly PM10) were collected in the valley. This field campaign along with preliminary results are presented in a companion paper by Paci et al.

The aim of the present paper is to introduce the objectives and underlying scientific questions of the project, and to report about preliminary numerical experiments of the atmospheric circulation in the valley using the WRF model. Nested domains are used for this purpose but such a numerical modelling represents a technical challenge: the valley around Passy is very narrow (few kilometres in width) with steep slopes. Numerical results for the diurnal cycle of a cold-air pool during one of the IOPs will be compared to field measurements, focusing on the height of the inversion layer and on wind profiles. A preliminary picture of the atmospheric circulation in the valley will be proposed, the attention being brought to the ventilation processes inside the cold-air pool.

High-resolution numerical simulations are performed with the Weather Research and Forecasting (WRF) model and with the CALPUFF dispersion model in order to reproduce meteorological conditions and particulate concentration vertical profiles close to the city of Merano, in the Adige Valley, in the north-eastern Italian Alps. Simulation results are compared against observations recorded both by surface weather stations and by means of balloon soundings performed during a field campaign in March 2010. The soundings recorded temperature and particulate concentrations from the surface up to 800 m above ground level in several moments of the day, to evaluate the temporal evolution of the boundary layer. Interesting meteorological phenomena, able to influence the dispersion of particulate matter, were observed during the field campaign and were also correctly reproduced by the simulations, as for instance the break-up of the early morning ground-based thermal inversion after sunrise and the arrival of strong foehn winds. The observations clearly show that these meteorological phenomena are strictly related with changes in the vertical distribution of particulate matter: concentrations, indeed, are very high near the ground in the early morning, when thermal inversion, biomass burning for domestic heating and the traffic rush hour act simultaneously. Conversely, particulate matter concentrations diluted with the development of the boundary-layer and strongly decreased at all heights when the foehn wind starts to blow.

Dispersion modelling in an Alpine surrounding has to cope with steep slopes, prolonged periods with low wind speeds and cold air drainage during night-time. Diagnostic wind fields have clear limits in dealing with these aspects, while complex models like WRF (Grell et al. 2005) and INCA (Haiden et al. 2010) should render more realistic wind fields under these conditions. The research-project WAAR focused on the question, which model is to favour and to quantify the impact on next-in-line dispersion models.

The basic dataset used for this evaluation was provided by instruments run in the framework of the “Innsbruck Box” (Stiperski et al. 2012) by the Institute for Meteorology and Geophysics, University of Innsbruck (IMGI). Data from 5 different sites (valley floor and adjacent slopes) around Kolsass (Inn-Valley) were available.

Half hourly wind fields were calculated for two episodes with low windspeeds and two episodes with stronger wind (one foehn event, one frontal passage) with LASAT (Janicke, 2013) and GRAMM (Oettl et al., 2010) , both using the valley-floor data as input and a horizontal resolution of 50 m.

Furthermore the selected episodes were simulated (without taking the locally measured data into account) with the mesoscale prediction model WRF (Grell et al. 2005) and the Analysis and Nowcasting system INCA developed at ZAMG (Haiden et al. 2010) on an hourly basis. INCA is based on the mesoscale prognostic model ALARO; meteorological parameters (i.e. wind, temperature, humidity, precipitation, cloud cover) are provided hourly with a horizontal resolution of 1 km. WRF-CHEM is based on three-hourly data (analysis for 0 and 12 UTC; prognosis for the rest of the time) from the ECMWF-model. The operational grid of WRF-CHEM (Alpine area) has a horizontal resolution of 4 km. Within this grid a nest (100x100 km) with a horizontal resolution of 1 km was centred above the project area. In order to meet the needs of dispersion modelling (LASAT) on a local scale the wind fields simulated by INCA and WRF had to be downscaled (using a bilinear interpolation scheme) to a horizontal resolution of 50 m. These interpolated windfields were used as additional input for LASAT.

Results:

Along the slopes INCA and WRF deliver better results. On the valley-floor (near the measurement site) LASAT and GRAMM/GRAL produce more realistic wind fields.

Against expectations INCA and WRF did not reproduce the two episodes with strong winds very well. There is still some need for an improvement of the interpolation-algorithms, especially for wind speeds in WRF.

For the foehn-event the wind fields are discussed in terms of meteorological plausibility on the valley bottom and on the slopes. A comparison with the i-Box stations and statistical analysis is presented as well and serves to validate the results. The effects on a typical dispersion calculation (i.e. a highway along the valley) are shown qualitatively.

The 2014 fissure eruption in Holuhraun is unique among recent eruptions in Iceland for its high emission rates of volcanic gases. The plume was relatively ash free, but predominantly a bent over vapour plume and its height depended mainly on the atmospheric conditions at the eruption site.

During the first month and a half the preliminary SO2 flux was ~400 kg/s with some days greater than 1000 kg/s. The gas is dispersed from the eruption and was transported by wind, leading to high pollution levels in exposed populated areas in Iceland. During high wind events and when nearby weather systems led to rapid change in wind directions the local population was not been much affected by the emission.

However, during certain conditions, usually light winds and low-level temperature inversions, the concentration of gas build up at the eruption site and then either flowed down from the highlands with katabatic wind or was advected from the eruption site when the synoptic situation changes. Depending on the atmospheric conditions, high concentrations of SO2 were transported in the boundary layer and detected at ground level in populated areas.

Here we describe one such event, the event of 26 and 27 October 2014, when the village Höfn, in southeast-Iceland, experienced gas concentrations exceeding 14000 µg/m3, a concentration considered hazardous to health. We describe the weather conditions prior and during the event as well as the gas dispersion.

Important ecosystem functions, such as the pools and exchange fluxes of carbon, nitrogen, energy and water, or soil fertility, are expected to be altered substantially due to global environmental change. This applies especially for mountain regions which are among the most climate change-sensitive landscapes. The manifold spatial complexity of mountain regions involves variations of terrain, soil types, water availability, vegetation, and land-use across a wide spectrum of scales. The spatial complexity is a challenge for both observation and modeling of ecosystem-atmosphere interactions, boundary layer dynamics, patterns of atmospheric circulation, precipitation, subsurface water movement and nutrient flows.

TERENO is a large-scale integrative observation program designed to determine the long-term ecological and climatic impact of global environmental change at regional scales. The TERENO-prealpine observatory in southern Germany comprises three core grassland sites at different elevations (600, 750 and 860 m a.s.l), which are equipped with eddy-covariance flux towers and chamber-flux/lysimeter clusters. The climasequence arrangement of the core sites along an elevation gradient serves as an in-situ climate change analogy. The complex environment poses a challenging research question: how well can our observations constrain modeling uncertainties of biogeochemical cycles, and close the balances of energy and matter flows? This research question integrates one-dimensional single site approaches, transects along climatic and terrain gradients, and spatially explicit three-dimensional representations of land surface – atmosphere interactions.

The ScaleX campaign in the TERENO-prealpine observatroy is designed to address this overarching research question, and to expand our understanding of energy and matter fluxes from site to regional scale. The scale expansion will be achieved by linking observations of surface fluxes (lysimeter, chamber, EC) to remote sensing techniques (LIDAR, Sodar-RASS, digital and hyperspectral imagery, X-band radar), airborne measurements (ultralight aircraft and unmanned aerial vehicle), and process- to meso-scale modeling (WRF-hydro, WRF-CHEM, LES, L-DNDC). ScaleX is structured into work-packages to examine specific aspects of land surface-atmosphere interactions, including:

• Effects of mesoscale circulations on biosphere-atmosphere exchange processes

• Trace gas budgets in the nocturnal boundary layer for mesoscale model evaluation

• Patterns of precipitation and soil moisture from site to regional scale

• Capability of digital/hyperspectral imagery to scale canopy traits of ecosystem function

• Closure of atmospheric water and energy cycles

• Distributed modeling of energy, water, carbon and nutrient cycles from local to regional scales

We will present preliminary results from the first phase of ScaleX, which will take place during June and July 2015. We will also give an outlook to the second phase, planned for 2016.

The planar fit method is often used for long-term eddy-covariance flux measurements since it offers a number of advantages over rotating each 15-minute sample into streamwise coordinates. For sites over complex, forested terrain however a single planar fit may not be appropriate since such sites may have very different slopes and forest cover in different directions. There is significant interest in measuring fluxes of momentum, heat and other scalars at such sites to understand the role of complex topography and heterogeneous forest canopies on surface exchange.

We present an alternative to the standard planar fit method where the tilt angle is fitted as a continuous function of the wind direction. This retains many of the benefits of the planar fit method, while at the same time better representing local variations in tilt angle with wind direction. The method is tested for momentum and heat fluxes using data from a field experiment on the Isle of Arran, Scotland studying forest canopy-atmosphere interactions over complex terrain. The 3 tower sites are located over a ridge and there is significant heterogeneity in the canopy cover around them. This makes it a tough test for any coordinate transformation scheme. Results are compared with the traditional planar fit, streamwise rotation and also a sector planar fit method. In most cases the new continuous planar fit method demonstrates better agreement with fluxes obtained from rotating into streamwise coordinates when compared to the standard planar fit method. It also behaves more smoothly than the sector planar fit method, avoiding the discontinuities which can be seen using this method.

Measurement and analysis of turbulence in mountainous terrain are challenging in several respects. Non-ideal terrain makes finding an appropriate coordinate system non-trivial, whereas the abundance of motions on all scales in mountainous terrain raises the question of isolating true turbulence. Still the post-processing methods for analysing turbulence statistics are based to a large extent on the assumptions of horizontally homogeneous and flat terrain.

Here we discuss those aspects of turbulence measurements that require special attention in complex mountainous terrain. We place special focus on coordinate systems and different post-processing options in mountainous terrain. The appropriate choice of post-processing methods is then tested based on local scaling arguments. It is shown that over non-ideal slopes the currently commonly applied planar fit method can be inappropriate, lead to erroneous scaling and underestimates the flux. The reasons for this are studied in detail and a modified planar fit method suggested. We also demonstrate that conclusions drawn from turbulence measurements in complex terrain (e.g., scaling relations) are sensitive to these post-processing choices, thus questioning the results obtained in other studies so far.

Due to the development (or presence) of internal boundary layers the turbulent characteristics in the lowest meters over heterogeneous surfaces are more complex than above horizontally homogeneous and flat terrain. In this study we investigate the applicability of Nieuwstadt's (1984) local scaling formulation for the stable boundary layer. The data for this study stem from a tall tower within a small forest patch situated in heterogeneous terrain with agricultural land, forested hills and urban surfaces in different upwind sectors. The 62 m tower (levels 20, 32, 40, 55 and 62 m above ground) was situated in the middle of some 120 m x 480 m area of 18 m high trees. Here we only analyze periods with stable stratification from winter 2008/2009 focusing on the influence of the upwind hilly terrain. As turbulent fluxes showed a substantial variation with height, we adopt local scaling approach for which similarity functions and the local stability parameter are based on local fluxes at measurement height. In the data analysis the role of self-correlation is examined. Also an in-depth error analysis applying the Filtering Method allows to estimate the percentage of random errors in the turbulence variables. Using this method it is possible to inspect if the scatter in the measurements is caused by random errors, or by other dynamical factors of the stable boundary layer. Values of scaled standard deviation for wind components in near-neutral conditions are found to be lower at the lowest measurement level and higher at upper levels in comparison to canonical Kansas values. The non-dimensional gradient of wind velocity is also investigated. We compare our results with traditional linear equations for the stable case and also to the empirical non-linear expression proposed by Beljaars and Holtslag (1991). It is found that the stability function for momentum supports the linear equation only up to values of the non-dimensional stability parameter ζ ≈ 0.5. Moreover, we find good agreement between our results and the Beljaars and Holtslag function, which increases more slowly with increasing stability. As we observed more strongly stable stationary cases, local z-less scaling is also addressed. Our analysis supports the validity of z-less stratification for very stable conditions (ζ > 1) for scaled wind velocity variances and correlation coefficients for momentum and heat flux. As a preliminary conclusion, based on scaled velocity variances and non-dimensional mean wind shear, it is found that local scaling is promising even over highly non-homogeneous terrain as in our case.

At the majority of FLUXNET sites the net ecosystem exchanges of energy and carbon dioxide are quantified based on a simplified, one-dimensional mass balance, which neglects advective flux contributions. In particular in complex terrain and nighttime conditions, this approach is questionable and thought to cause an underestimation of the true net ecosystem exchange. For carbon dioxide, which typically is taken up during daytime and released during nighttime, this is likely to cause a systematic bias.

Here we report on an experiment at a subalpine mountain grassland in Northern Italy situated in truly complex terrain characterized by slopes of different inclination and exposition. During the month-long experiment we attempted to quantify all terms of the full three-dimensional carbon dioxide mass balance and compared these with concurrent ecosystem respiration measurements by automated chambers.

The main findings of this study can be summarized as follows: (i) the sum of the vertical covariance term and the storage term considerably underestimates nighttime ecosystem respiration as measured by the automated ecosystem chambers; (ii) advection measurements indicate that both horizontal and (less so) vertical advection are important terms of the full mass balance during nighttime; (iii) the net ecosystem carbon dioxide exchange calculated by taking into account all terms, i.e. including advection, closely resembles nighttime ecosystem respiration as measured with the automated ecosystem chambers; (iv) there is substantial spatial variability in the vertical covariance term during nighttime; (v) during daytime, advection appears to make a negligible contribution to net ecosystem carbon dioxide exchange.

Sensible heat fluxes from the ground to the atmosphere cause the planetary boundary layer to become convectively mixed. The spatial variability of mixing heights is known to be determined both by the distribution of heat fluxes and by the shape of the underlying topography, mostly through two distinct physical processes. First, heat fluxes that are either relatively strong or occur at relatively high altitude, e.g. at mountain tops, locally enhance vertical mixing and cause the convective boundary layer (CBL) depth to increase. In addition, the mixing height can be modulated by the lateral redistribution of air masses and heat by breeze systems, generated both by differential heat fluxes and by topographic forcing.

Little is known on which of these factors, land-use differences and topography, has the largest potential impact on mixing height variability. We address this issue by performing a quantitative case study on the area surrounding Granite Peak, in north-western Utah. Granite Peak is an isolated mountain rising approximately 800 m above the surrounding terrain. It separates a flat dry lake (playa) to the west from arid shrubland to the east. The plain east of the peak slopes gently towards the northwest. Upslope winds along the sidewalls of the mountain and lake breezes between the playa and the adjacent plain develop on fair weather days with strong solar forcing.

We perform semi-idealized very-large eddy simulations and use a factor separation approach to assess the pure impacts of topographic effects and of the spatial heterogeneity of land-use on CBL depth. We show that, when synoptic forcing is weak, mixing height variability in this area depends mostly on topographic effects. Beyond quiescent environmental conditions, we examine different prototypical flows with large-scale winds impinging on Granite Peak from the south and the north. Finally, we attempt to quantify how the balance between topographical and land-use effects changes in response to variable strengths of the two factors.

The Second Meteor Crater Experiment (METCRAX II) was designed to study downslope-windstorm-type flows (DWF) that occur above the south and west inner sidewalls of the 1.2 km diameter Barringer Meteorite Crater in Arizona. These DWFs occur intermittently within the crater basin on clear, undisturbed nights in connection with a mesoscale drainage flow that approaches the crater from higher terrain to the southwest of the crater. During DWF events a wave descends in the lee of the upwind crater rim, producing a strong and turbulent downslope flow above the slope, which rebounds in a hydraulic jump-like flow feature at the base of the slope. Multiple break-ins of these flows occur on suitable nights, and the individual break-ins vary in strength and duration.

This presentation will provide an overview of the October 2013 METCRAX II field experiment and summarize recent research results. The experiment was supported by the National Center for Atmospheric Research's Earth Observing Laboratory and by other organizations, with an extensive complement of field equipment including a radar wind profiler, 2 radio acoustic sounding systems, three scanning Doppler LiDARs, a vertically pointing LiDAR, instrumented 40- and 50-m towers, 3 surface flux stations, 3 SoDARs, a scintillometer, a ceilometer, automatic weather stations, an array of temperature data loggers, multiple pressure sensors, and visual and thermal IR time-lapse cameras. Seven overnight Intensive Observational Periods (IOPs) were selected during the month on the basis of weather forecasts. Three continuously operated tethersondes and 3-hourly radiosondes provided supplementary vertical sounding data during the IOPs. Forecasts were successful in selecting nights when the DWF phenomenon occurred, and analyses of the extensive data are now in progress. This presentation will summarize recent results, leaving detailed analyses for presentation by other team members in related oral and poster presentations at the conference.

The second Meteor Crater Experiment (METCRAX II) field campaign was conducted at Arizona's Meteor Crater in October 2013 to study the regular nighttime occurrence of downslope-windstorm-type flows (DWFs) in the crater basin. The almost circular and approximately 1.2-km wide Meteor Crater is located on a slightly sloping plain, where a southwesterly katabatic flow forms during undisturbed, clear-sky nights. As the southwesterly flow approaches the Meteor Crater, cold air pools upstream of the 30-50-m high crater rim and partially drains into the crater in the form of cold-air intrusions. While part of the katabatic flow diverges around the crater and upstream cold-air pool, part of it flows over the crater, leading to the formation of DWFs along the inner southwest sidewall.

In this presentation, we will look at the response of the basin atmosphere to the upstream katabatic flow that impinges on the crater topography, including the formation of DWFs along the sidewall and the occurrence of seiches in the crater cold-air pool. It is hypothesized that the seiches observed in the surface-based inversion on the crater floor are related to oscillations in the cold-air intrusions down the southwest sidewall and to oscillations in the upstream katabatic flow. On the basis of observations and numerical simulations, variations in the strength and depth of the upstream katabatic flow are identified as key factors that determine the formation and strength of DWFs in the crater basin.

The second Meteor Crater Experiment (METCRAX-II) was designed to study downslope-windstorm type flows in the lee of the upwind crater rim of Arizona's Meteor Crater. These flows are frequently observed during synoptically quiescent nights when a southwesterly mesoscale drainage flow forming outside of the crater basin interacts with the crater topography.

Three lidars were deployed during METCRAX-II. One lidar was positioned on the upwind side of the crater and was used to monitor and characterize the mesoscale drainage flow approaching the crater. The two other lidars, one on the crater's northern rim and one on the floor of the crater, were positioned on a NNE-SSW transect across the crater topography. Co-planar range-height indicator scans along this transect allowed the retrieval of the two-dimensional wind field of the intruding flows. Flow features such as hydraulic jumps, rotors and flow separation can be resolved from the dataset.

We will show the temporal evolution of the drainage flow outside of the crater topography and its flow response to the topographic obstacle, including flow stagnation near the surface and flow splitting around the crater. Further, we will illustrate the development and evolution of the downslope-windstorm-type intrusions into the crater basin, from shallow cold-air inflows along the inner crater sidewalls to flow separation and the formation of hydraulic jumps, rotors and return circulations.

Observations gathered during the METCRAX 2 field campaign show that the incoming flow associated with downslope-windstorm-type flows (DWFs) within Arizona’s Barringer Meteor Crater is highly variable in time and space. Such flow complexity can make analysis problematic, with the chaotic structure potentially masking important flow features. In order to improve our understanding of DWFs, we create idealised and semi-idealised simulations of the Meteor Crater environment. These allow us to better understand the flow features that control DWF generation, and their internal dynamics and structure.

Using METCRAX 2 observations as a starting point for the atmospheric structure in our model, fully nonlinear 2D simulations using the real crater terrain are presented. Using a high resolution grid in both horizontal (10 m) and vertical (5 m) directions, we are able to fully resolve the generation of the plunging air into the crater, and the dynamical structure of the wave-like response within the crater atmosphere.

With our approach, we are able to specify the structure and location of the flow features of interest, allowing us to directly determine the impact of an upstream change on the resulting flow. In this manner, we study parameters such as the importance of the speed of the incoming flow, its stability, and its depth and location relative to the crater rim.

Initial simulations show that we can simulate a DWF and warm air intrusion of comparable strength to those in the observations. Associated with this is evidence of boundary layer separation, a hydraulic jump, and a standing wave train - all features observed using LIDAR during METCRAX 2. From our initial work the strength of the cold pool within the Meteor Crater appears unrelated to the generation of DWFs.

Theories of nonlinear lee-side flow amplification have been developed to explain severe windstorms observed downstream of large mountain ranges where the drop in elevation is typically 1-2 km. Observations during METCRAX II of (much weaker) downslope-windstorm-type flows (DWFs) in the Meteor Crater, which is just 150 m deep, raises the question whether existing theory is able to account for such flows as well. One of the main differences, apart from the smaller-scale and curved topography, is the fact that the flow is embedded entirely within the atmospheric boundary-layer, which increases the relative importance of surface friction. Due to the smaller horizontal scale also non-hydrostatic effects may be important. Based on the comprehensive dataset obtained during the METCRAX II field experiment we address these questions by estimating individual terms in the momentum equation. The presence in the crater of a flow feature resembling a hydraulic jump is analysed with reference to the theory of continuously stratified hydraulic flow and the theory of wave amplification in a flow with a stable layer near the surface. It is analyzed to what extent these theories are able to predict the occurrence of DWFs in the crater as a function of upstream flow characteristics.

During clear and quiescent nights downslope-windstorm-type flows associated with warm air intrusions often occur in the Barringer Meteor Crater in northern Arizona. Their occurrence presumably strongly depends on the upstream flow conditions, particularly on the depth of the mesoscale drainage flow, which regularly forms on the low-angle slope surrounding the Meteor Crater basin. During the second Meteor Crater Experiment (METCRAX II) performed in autumn 2013, comprehensive in-situ and remote sensing instruments were installed inside the crater, on the crater rim and on the slope surrounding the crater.

We investigated the relation between the upstream flow conditions (stability, strength and depth of the mesoscale drainage flow) and the conditions above and inside the crater basin as well as downstream of the crater. Depending on the depth of the upstream mesoscale drainage flow, the flow responses inside and downstream of the crater topography varied When the upstream mesoscale drainage flow was shallow, i.e. about 50 m deep, parts of it intruded into the crater basin along the upstream slope as cold air intrusion. This resulted in an outflow of air over the downstream crater rim. In these cases, no mesoscale drainage flow was observed above or downstream of the crater basin. During some nights the upstream mesoscale drainage flow reached depths of up to 200 m. This was associated with downslope-windstorm-type flows and high turbulence inside the crater. In these cases a flow with similar characteristics as the upstream mesoscale drainage flow was observed downstream of the crater basin.

Observations and physical measurements have traditionally played a pivotal role in complex terrain research. In order to advance understanding of physical processes and phenomena in complex terrain, motivated in large part by an ongoing need to improve forecasting of weather in complex terrain, this field of atmospheric research has been at the forefront in adoption of new technologies and measurement techniques as well as in novel applications of the existing ones.

In this presentation we will offer a retrospective view of some of the key advances in complex terrain research that had been made possible by the technological advances of the day and make an attempt at identifying remaining challenges. Special emphasis will be placed on progress that has been achieved within the context of large observational field campaigns.

Gravity waves play an important role in regulating the momentum budget of the global circulation as they propagate vertically from source regions in the troposphere through the stratosphere and into the mesosphere. Therefore, it is critical to understand the predictive capability of gravity waves in numerical weather prediction. In this talk we will examine the predictability of orographic and non-orographic gravity waves generated over the South Island of New Zealand and Tasman Sea during the DEEPWAVE field campaign. The Coupled Ocean/Atmosphere Mesoscale Prediction System (COAMPS) will be used in an ensemble mode to diagnose the characteristics of initial condition perturbation growth in gravity waves as they transit from the troposphere into the stratosphere. Initial results show that compared to the troposphere, perturbation growth rates for orographic gravity waves are suppressed in the stratosphere, and that gravity waves may experience enhanced predictability there. The mechanisms for the decreased growth rates will be explored. The predictability of non-orographic gravity waves in the troposphere and stratosphere will also be presented.

The Deep Propagating Gravity Wave Experiment (DEEPWAVE) field campaign was conducted over the New Zealand region to study gravity waves from tropospheric origins to their dissipation at very high altitudes. Preliminary analysis from a season long 6-km resolution WRF simulation suggests about half of New Zealand mountain wave events are strongly attenuated in a lower stratosphere “valve layer” near 15 km. How are mountain waves attenuated in this layer? How are waves modified as they pass through this layer? Does mountain wave attenuation violate PV conservation? These questions are investigated with NCAR Gulfstream V (NGV) aircraft observations and high-resolution simulations of such a valve layer attenuation event that occurred on 24 May 2014 over the South Island of New Zealand. The NGV provided in situ measurements at 12 and 13.5 km and sampled wave breaking along higher flight legs. A microwave temperature profiler onboard the NGV provided remotely sensed 2-D cross sections of temperature 5 km above and below the aircraft. High-resolution (2-km) WRF simulations are compared with the aircraft and AIRS observations and used to study the dynamics of propagation and attenuation through the valve layer. A new 2-D spectral filtering method is used to quantify resolved gravity wave momentum flux and drag within the simulated fields, which suggest the strongest attenuation occurred near 14 km resulting in gravity wave drag of roughly 25 m s-1 day-1.

An almost linear theory of mountain gravity waves present when the incident wind a is (1) null at the ground and upstream the ridge and (2) increasing with altitude is presented. It partly solves the fundamental problem of the “classical” linear mountain wave theories, where the surface winds need to be non-zero at the surface for trapped mountain waves to develop substantially. In these theories, trapped lee-waves occur when there are turning points aloft and providing that the waves reflected toward the surface by these turning points are not absorbed in the boundary layer. These “classical” theories do not take into account that in reality the winds go to zero at the surface, a conceptual difficulty because a zero surface wind corresponds to a critical level for mountain waves, that is a place where gravity waves absorption can be very strong, at least when the flow is stable according to the Miles-Howard criteria. We therefore show that trapped lee waves are favoured when the surface Richardson number is small, bridging a conceptual relation between trapped lee waves and Kelvin Helmholtz instabilities.

We hope that the solutions presented here could help to benchmark Numerical Weather Prediction models that aim at representing mountains and including the smallest scales gravity waves. The solutions proposed can also be used to analyse at a cheap numerical cost the refraction by the gravity waves and the mountains of the infrasounds trapped in the low tropospheric sonic waveguide.

Some details about the mathematical functions used can be found in:

F. Lott, C. Millet, and J. Vanneste, 2014: Inertia-gravity waves in inertially stable and unstable sheared flows, Submitted to J. Fluid Mech.

Lott, F., The reflection of a stationary gravity wave by a viscous boundary layer, 2007: Journal of the Atmospheric Sciences, 64, 3363–3371.

The South Georgia wave experiment (SG-WEX) combines numerical modelling and observations from radiosonde, meteor radar and remote sensing to investigate gravity wave activity in the vicinity of South Georgia. Located in the South Atlantic, downstream of the Drake passage and collocated with a peak in stratospheric gravity wave activity the island of South Georgia provides an ideal natural laboratory to address key uncertainties in our understanding of the contribution of small islands to gravity wave fluxes and low level drag.

Results from simulations completed as part of SG-WEX will be discussed, with particular focus on validation of the low level flow around South Georgia. Simulations have been completed using the Met Office NWP model in a nested domain configuration centred on South Georgia at 1.5km horizontal resolution and enhanced vertical domain (78km). At this resolution the model is able to resolve gravity wave activity and the low level flow structure around the island. These simulations are then compared against lower resolution global NWP configurations which use an orographic parameterisation scheme to represent the gravity wave and low level drag. The high resolution simulation results are compared with ASCAT scatterometer data of the surface wind structure in the wake and with radiosonde data from the SG-WEX field campaigns in order to assess the performance of the model. Using ASCAT dataset combined with ECMWF ERA-Interim reanalysis a climatology of low level flow regimes around South Georgia is created in order to assess the likely impact of South Georgia on the regional and zonal flow.

The instability of the eastern Antarctic Peninsula (AP) ice shelves have become a symbol for climate change in the Polar Regions. The northernmost sections of the Larsen Ice Shelf (LIS) collapsed in 1995 and 2002, prompting research into the cause of the disintegration. One theory behind this instability is the ‘föhn hypothesis’. The föhn winds which flow down the eastern slopes of the AP are a feature of the interaction of the mountain range with the prevailing circumpolar westerlies. The theory suggests that the warm, dry conditions which characterise föhn winds are prompting surface melting of the LIS. Investigating the spatial distribution of these föhn winds, and the impact they are having on the remaining Larsen C ice shelf formed part of the NERC funded project ‘Orographic Flows and the Climate of the Antarctic Peninsula’ (OFCAP). As part of this campaign an automatic weather station was installed on Cole Peninsula (-66°51'S, -63°48'W). This station, along with six others provides a network of observational instruments which is complemented with data from the Antarctic Mesoscale Prediction System (AMPS) archive. The AMPS archive holds outputs from the Weather Research and Forecast (WRF) model run operationally over the Antarctic and Southern ocean by the National Centre for Atmospheric Research, USA. Archived model output covering the AP (domain six) from January 1st 2009 to January 1st 2012 is used here.

The occurrence and characteristics of föhn conditions over the LIS, identified from a combination of observational and model data, will be presented here. A north-south and west-east gradient in strength and occurrence of föhn events is apparent, as are other localised spatial variations in conditions. A comparison of the föhn characteristics within the observational and model data shows that there is a good overall agreement between the two datasets. However, there are variations in timing, strength and frequency of the föhn conditions within the data. A number of case studies will also be presented, which highlight the variance in föhn conditions across the shelf.

In Alpine mountain meteorology, orographic blocking is important because of its impact on: a) lee-cyclogenesis; b) the passage and modification of warm and cold fronts; c) the geographical distribution and intensity of (heavy) precipitation events. Traditionally, orographic blocking is studied with Eulerian methods based on the estimation of the inverse Froude number. Here we present a Lagrangian perspective of orographic blocking, and focus in particular on its impact on precipitation patterns.

Winds are taken from a three-year (2000-2002) reanalysis simulation with COSMO at a horizontal resolution of 7 km. Based on these winds, kinematic forward trajectories are started at a distance of 300 km all around the Alps and at two height levels (750 and 1500 m). The 24-h trajectories are then investigated in their capability to surmount the Alpine barrier, separated into three distinct flow categories, also defined in a Lagrangian way: westerly flow, northerly flow and southerly flow, the latter being restricted to south Foehn cases. As a result a Lagrangian blocking index is available at an hourly resolution between 2000-2002.

For each flow category the percentage of trajectories surmounting the Alps and the percentage of air parcels flowing around the Alps is determined. Furthermore, trajectory densities are calculated to show the different air streams which start from selected upstream positions. Composites of precipitation pattern are shown according to flow category and Lagrangian blocking index.

Finally, an outlook will be given to a new project based on high-resolution COSMO simulations. They will cover whole of Europe at 2-km horizontal resolution during a 10-year time period. A first glimpse on orographic blocking in this data set will be provided.

The capabilities and complexities of weather forecast models have increased greatly since they were introduced in the 1950s. Operational models are now capable of running at O(1 km) resolution, which means the steepness of the terrain that can be represented has increased dramatically. On the urban scale, models are now capable of simulating flow around buildings at O(1 m) resolution. With any numerical simulation, it is important to understand the role of numerical errors on the solution. These can be discretization errors related to the grid and the numerical methods chosen, or errors from the physical parameterization approaches chosen to represent unresolved turbulence or land-surface forcing, for example. For flow over complex terrain, additional complications can arise. We will discuss errors due to the choice of coordinate system and grid aspect ratio, alternatives to terrain-following coordinates such as the immersed boundary method (IBM), and grid nesting strategies. The importance of physical parameterizations will be emphasized through discussion of turbulence closure models and land-surface models, e.g. illustrating the role of soil moisture variability on flow over complex terrain. A range of domain sizes and grid resolutions will be discussed, encompassing the Reynolds-averaged Navier-Stokes (RANS) and large-eddy simulation (LES) techniques, and the gray zone, where the model resolution is in the intermediate range between LES and RANS.

Future very-high resolution mesoscale weather prediction models will require a robust and efficient dynamical core allowing for explicit representation of vigorous convective processes involving close coupling of dynamics and physics, as well as successful handling of steep mountain slopes. Such a core should represent also the basic conservative properties of natural flows.

New EULAG dynamical core for the operational Consortium for Small-scale Modeling (COSMO) model is currently being developed at IMGW within the CELO priority project of COSMO. Recently, its forecasting skills were compared against the reference operational Runge-Kutta dynamical core, within the COSMO framework set close to the operational 2.2 km horizontal resolution setup over the Alps. Within this presentation, an overview of the new dynamical core capabilities, highlights of the verification scores and the projected performance will be presented, including studies at the very high horizontal resolutions up to 100 m.

As we move toward multi-scale modeling, vertical grid nesting becomes necessary for simulating a wider range of scales. This talk presents a new vertical grid nesting capability we have implemented in the Weather Research and Forecasting (WRF) model based on the interpolation method used by Moustaoui et al. 2007. Vertical grid nesting potentially lowers computational expenses and allows control over grid aspect ratio for improved model stability and accuracy. Grid aspect ratio can be important for reducing errors in large-eddy simulations (Mirocha et al. 2013) and for mesoscale simulations over complex terrain where steep terrain slopes can cause errors in the calculation of horizontal gradients. Idealized simulations are presented which validate the vertical nesting method by showing that solutions from vertically nested and non-nested cases compare well. Large-eddy simulations of the atmospheric boundary layer over flat terrain are presented to demonstrate use of vertical nesting to control grid aspect ratio and improve simulation results when compared to the theoretical log-law solution. Mesoscale simulations over Owens Valley, California are presented and results are compared to observations from the Terrain-Induced Rotor Experiment (T-REX, March-April 2006).

References:

Mirocha, J., G. Kirkil, E. Bou-Zeid, F. K. Chow, and B. Kosovic, 2013: Transition and Equilibration of Neutral Atmospheric Boundary Layer Flow in One-Way Nested Large-Eddy Simulations Using the Weather Research and Forecasting Model. Monthly Weather Review, 141 (3), 918– 940.

Moustaoui, M., A. Mahalov, and B. Nicols, 2007: WRF and vertical nesting: multi-scale resolution of T-REX measurements. 8th WRF Users’ Workshop, Boulder, CO, National Center for Atmospheric Research, 1–8.

Assessing the local momentum budget is important for understanding atmospheric variability on a range of timescales, including the circulation response to climate change. However, the ability of general circulation models (GCMs) to correctly represent the momentum budget is poorly understood, and model circulation is sensitive to the parametrization of drag processes. These drag parametrization schemes are themselves poorly constrained. For example, the partition between orographic drag (due to mountain waves, and low-level flow blocking) and drag due to boundary-layer turbulence is tuneable and varies widely between different models. Furthermore, as the resolution of GCMs increases the drag processes are becoming increasingly (at least partially) resolved. With grid lengths of a few tens of kilometres, we are entering the "grey zone" of orographic drag parametrization. An understanding of how well parametrization schemes can represent the drag, and how well the hand-over between parametrized and resolved drag works across model resolutions is clearly important.

Recent studies have demonstrated how GCMs may also suffer from the so-called "small island problem", whereby the effect of small mountainous islands is neither resolved or parametrized (since they occupy too small a fraction of the GCM grid box). The importance of this problem is unclear, but recent work has suggested that small mountainous islands in the South Atlantic may make significant contributions to the stratospheric gravity-wave momentum flux and that this missing drag may result in systematic errors in the model southern hemisphere stratospheric circulation.

The focus of this presentation is the ability of orographic drag parametrization schemes to represent the drag associated with mountain waves and low-level flow-blocking processes. High resolution numerical simulations of flows over New Zealand and South Georgia Island will be used to determine the drag on the flow, and where possible these simulations will be compared with recent observations from the DEEPWAVE and SG-WEX field campaigns, respectively. The high resolution results will be compared with those from coarse resolution experiments in which the drag is largely parametrized. By varying the horizontal grid length, we will determine the extent to which the total (resolved plus parametrized) drag is invariant across model resolutions. By considering both relatively wide (New Zealand) and narrow (South Georgia) mountain ranges, the importance of mountain width relative to the grid length will be investigated, an issue related to the "small island" problem.

Recent advances in understanding and modelling snow processes in mountain environments are leading to better assessment of snow accumulation and ablation, and snow-atmosphere energy and mass exchange. One particularly important area of advance is in the coupling of complex terrain windflow with solar irradiance, wind transport, sublimation and melt calculations – this coupling is leading to detailed representations of snow accumulation, turbulent exchange with the atmosphere, albedo and longwave radiation emission in complex terrain. Results from this show that detailed physically based wind flow calculations are necessary to estimate the spatial patterns of alpine snow accumulation and ablation and that these spatial patterns control the areal depletion of snowcover and the persistence of ecologically and hydrologically important snow patches into summer. However, serious questions remain on how to apply stability corrections to turbulent fluxes in complex terrain. A challenge will be to represent this spatial heterogeneity at an appropriate level of complexity in upscaled mass and energy balance calculations for land surface schemes. A second area of advance is in the measurement, understanding and simulation of snow-vegetation interactions at treelines, in shrub-tundra and in discontinuous forest canopies. This has required a better understanding of canopy radiative, aerodynamic and thermal properties and the complex interactions between the atmosphere, discontinuous plant canopies and underlying or adjacent snowcover. Results from these investigations are showing that simple 1-D representations of plant canopies cannot succeed in providing the radiative and aerodynamic exchanges needed for accurate snow ablation calculations and that fascinating relationships between snow accumulation patterns, solar and thermal irradiance and turbulent transfer develop in sparse canopies and dominate snow dynamics in these environments.

In the mountains of the Western US, shifts in the timing and magnitude of snow water equivalent (SWE) over the past century are well documented and attributed to climate warming, but the magnitude of sensitivity depends on elevation. We estimated the spatial distribution of SWE and its sensitivity to climate warming in the topographically complex, 1500 km2 Upper Deschutes River Basin, Oregon (USA), with a spatially distributed snowpack energy balance model forced by a gridded meteorological dataset. We further compared how these estimates differed across elevations depending on the choice of lapse rates used for spatial temperature interpolation and extrapolation. The 1/16o spatial-scale gridded meteorological forcing data was bias-corrected with PRISM climate data and downscaled to a 100-m spatial-scale digital elevation model using 1) a spatially uniform and temporally constant -6.5oC km-1 lapse rate, and 2) with monthly varying lapse rates computed from long term records of temperature recorded by weather stations in the study region. Model parameters that controlled empirical estimates of incoming shortwave and longwave irradiance and the partitioning of precipitation into rain and snow were estimated independently with each downscaled forcing dataset to optimize the agreement between modeled and observed SWE. We then estimated the sensitivity of the snowpack to +2oC and +4oC warming with each of the four downscaled temperature datasets and optimized parameters. Interdataset differences in modeled SWE during the historical period were largely driven by differences in estimated cloudy-day incoming longwave irradiance, that were in turn driven by differences in prescribed lapse rates. The sensitivity of SWE to +2oC and +4oC warming differed significantly at all elevations between the bias-corrected and original data, but did not depend on choice of lapse rates. At all elevations, SWE sensitivity was largely driven by shifts from snow to rain, but at high elevations increased estimates of longwave irradiance drove increased mid winter snowmelt. Our results revealed a previously unrecognized interaction between prescribed lapse rates and empirical estimates of incoming longwave irradiance, and demonstrate the challenges of modeling SWE with gridded data products in data sparse regions.

The low resolution and simplified processes used in climate models have long been issues for prediction of climate change impacts on snow in mountainous terrain. Upcoming phases of the Snow Model Intercomparison Project (ESM-SnowMIP), the Global Soil Wetness Project (GSWP3) and the Coupled Model Intercomparison Project (CMIP6) will, amongst many other variables, produce global snow simulations for longer periods and at higher resolutions than ever before. In preparation for ESM-SnowMIP, this presentation will review the performance of snow models driven with global driving data at 0.5 degree resolution, with and without downscaling, and in situ meteorological measurements from well-instrumented snow study sites in the Alps and North America.

Physically based models of the hydrological cycle are critical for testing our understanding of the natural world and enabling forecasting of extreme events. Previous intercomparison studies (i.e. SNOWMIP I & II, PILPS) of existing snow models that vary in complexity have been hampered by multiple differences in model structure. Recent efforts to encompass multiple model hypothesizes into a single framework (i.e. the Structure for Understanding Multiple Modeling Alternatives [SUMMA] model), have provided the tools necessary for a more rigorous validation of process representation. However, there exist few snow observatories that measure sufficient physical states and fluxes to fully constrain the possible combinations within these multiple model frameworks. In practice, observations of bulk snow states, such as the snow water equivalent (SWE) or snow depth, are most commonly available. The downfall of calibrating a snow model using such single bulk variables can lead to parameter equanimity and compensatory errors, which ultimately impacts the skill of a model as a predictive tool.

This study provides two examples of diagnosing modeled snow processes at a recently upgraded (Oct. 2012) snow study site located at Snoqualmie Pass (917 m), in the Washington Cascades, USA. We focused on two physical processes, new snow accumulation and snowpack outflow during mid-winter rain-on-snow events, for their importance towards controlling runoff and flooding in this rain-snow transitional basin. 40 years of historical snowboard measurements showed a mean 24hr accumulated snowfall density of 90 kg m-3 below 0°C, with a linear fraction of snow and rain between -1°C and +1°C wet-bulb temperature. Two years of observed snow pit temperature profiles from infrared cameras and manual thermometers found that cold biases in the model snowpack temperature prior to rain-on-snow events could delay model outflow multiple days compared to lysimeter observations. These two examples illustrate the utility of using multiple observations of internal snowpack states to constrain uncertain snow physics options.

Knowledge on changes in seasonal mountain snow water resources are essential e.g. for hydropower companies. Both, snow accumulation and ablation need to be investigated to make precise predictions of water stored in a seasonal snow cover. Only if the processes behind snow accumulation and ablation are understood with sufficient quantitative accuracy, the evolution of snow water resources under a changing climate can be addressed. It is known that different snow precipitation processes and snow redistribution are responsible for snow accumulation patterns in alpine terrain. In a small-scale analysis of radar data in the region of Davos, Mott et al. (2014) could identify different snow deposition patterns for homogeneous precipitation, seeder-feeder mechanism, preferential deposition and a combination of the seeder-feeder mechanism and preferential deposition. In addition to the snow precipitation mechanisms, snow redistribution due to snow-atmosphere interaction is essential to characterize snow accumulation patterns at small scales (Scipión et al., 2013).

In this study we investigate small-scale patterns of precipitation for an extended area over the Dischma valley (Davos, CH) for the winter season 2014/2015. An X-band polarimetric radar was installed on a slope facing the Dischma valley and it conducted plane position indicator (PPI) scans at elevation angles of 7° and 10° (minimum distance to the ground is about 300m and 500m, respectively) and three range height indicator (RHI) scans along the Dischma valley and along the Landwasser valley (i.e. along Davos). These radar products are available with horizontal and vertical resolution of 75 meters and a high temporal resolution of 5 minutes. The specific spatial patterns revealed by the radar measurements allow to characterize the different types of winter precipitation as well as to identify cloud microphysical and dynamical processes that govern the precipitation distribution. The continuous radar measurements are also used to analyze the frequency of certain types of hydrometeors and precipitation genesis processes as well as snow precipitation patters, which are related to specific atmospheric situations. For a few snowfall events, we additionally analyze terrestrial laser scans (TLS) of steep rock faces with different orientations that were performed before and after the snow precipitation events. The results allow us to relate identified accumulation patterns to the identified precipitation patterns and confirm the importance of redistribution processes for accumulation in steep terrain.

Mott, R., D. Scipión, M. Schneebeli, N. Dawes, A. Berne, and M. Lehning (2014), Orographic effects on snow deposition patterns in mountainous terrain, J. Geophys. Res. Atmos., 119, 1419-1439, doi:10.1002/2013JD019880.

Scipión, D.E., R. Mott, M.Lehning, M.Schneebeli, and A.Berne (2013), Seasonal small-scale spatial variability in alpine snowfall and snow accumulation, Water Resour. Res., 49, 1446-1457, doi:10.1002/wrcr.20135.

Atmospheric longwave radiation is a key component of the surface energy balance in mountainous catchments, influencing hydrologic process such as snowmelt, snow disappearance, and summer evapotranspiration. Many empirical parameterizations of atmospheric longwave radiation exist in the form of clear-sky emissivity and cloud correction models, and these are typically developed and calibrated at low elevation sites. In most mountainous catchments, it is not possible to evaluate the accuracy of these methods due to a lack of surface measurements. Here, we quantify uncertainty in longwave radiation due to parameterization choice and evaluate how uncertainty in longwave propagates to modeled snow and hydrologic processes. Based on 20 clear-sky emissivity models and 20 cloud correction methods, we examine an ensemble of 400 methods for calculating atmospheric longwave radiation and compare these to surface observations at three well-instrumented mountain research sites spanning maritime, intermountain, and continental climates. We compare and contrast how three physically based models of varying complexity (the Utah Energy Balance, SNOBAL, and the SHAW snow model) simulate snowpack development and ablation across the longwave methods. Using the SHAW model, we examine the linkages between summer evapotranspiration uncertainty and longwave uncertainty, and compare the relative contributions of growing season timing and length (as controlled by longwave-impacts on snow duration), energy availability (as a function of longwave parameterization), and water availability to evapotranspiration uncertainty. Initial results show that longwave uncertainty as a function of parameterization choice is large, with a range typically exceeding 100 Watts per square meter, and with significant impacts on modeled peak snow water equivalent (typically +/- 50%), mid-winter melt occurrence, surface energy feedbacks, and snow disappearance timing (1-2 months). The results highlight the need for robust parameterizations of longwave radiation for model-based assessments of snow and hydrologic processes, especially given the central role of longwave radiation to atmospheric warming associated with climate change.

Hydrological and energy balance models typically assume a continuous snow cover. Such approaches neglect that once the snow cover gets patchy in spring, additional boundary layer processes significantly affect the energy balance at the snow surface. Small-scale thermal internal boundary layers develop, involving strong vertical and horizontal flux divergences. Furthermore, the advection of warm air from bare ground towards snow-covered areas can promote strong atmospheric stabilities and boundary layer decoupling above snow that suppress net turbulent heat fluxes close to the snow surface. Neglecting these processes in energy balance models might result in a strong over- or underestimation of snow melt and runoff in spring.

We experimentally investigated the small-scale boundary layer dynamics over snow patches and their effect on the energy balance at the snow surface. A comprehensive measurement campaign, the Dischma Experiment, was conducted during the entire ablation periods of 2014 and 2015. The aim of this project is to investigate the boundary layer development and the energy exchange over a melting snow cover with a gradually decreasing snow cover fraction. Furthermore, the relative importance of different boundary layer processes is quantified on the local and catchments energy balance . For this purpose, two measurement towers equipped with five to six ultrasonic anemometers were installed along the wind fetch over a well-defined long-lasting snow patch. Furthermore, temporally and spatially high resolution ablation rates and snow surface temperatures were determined with a terrestrial laser scanner and an Infrared camera. This data set allows us to relate the spatial patterns of ablation rates and snow surface temperatures to boundary layer dynamics and the changing snow cover fraction. Experimental data reveal that wind velocity, wind fetch distance and topographical curvature control the boundary layer growth, boundary layer decoupling and the efficiency of advective heat transport to contribute to snow ablation. Data from a high density network of meteorological stations was further used to test numerical results from an atmospheric model that revealed a strong dependence of the relative importance of boundary layer processes and the development of local flow patterns on the snow patch size distribution and the synoptic wind forcing. We further verify previous wind tunnel experiments showing that boundary layer decoupling only evolves over snow patches located within a topographical depression and when ambient wind velocities are very low at the same time. The dependence of the advective heat flux and the potential occurrence of boundary layer decoupling to the snow cover fraction, synoptic wind forcing and the local topography will allow the parametrization and consideration of those small-scale effects in high-resolution energy balance models.

The ICON (ICOsahedral Nonhydrostatic) model has been developed since 2004 at the German Weather Service (DWD) and the Max-Planck-Institute for Meteorology (MPI-M) in order to achieve a unified modelling framework for global numerical weather prediction (NWP) and climate modelling. In 2012, the Karlsruhe Institute for Technology (KIT) joined the development team in order to incorporate the ART (Aerosols and Reactive Trace Gases) module, which is already coupled to the regional COSMO model. ICON became the operational global numerical weather prediction system at DWD on January 20, 2015, replacing the hydrostatic GME after about 15 years of operational production. Its main advantages over GME are exact local mass conservation, higher efficiency on massively parallel computer architectures, and more advanced physics paremterizations, which is one of the keys towards a better forecast quality.

This presentation will focus on the main improvmenents relevant for mountainous and/or snow-covered regions. From a numerical point of view, a major advantage of ICON over most other models with terrain-following coordinates is that it does not require orography filtering for numerical stability reasons. This allows restricting the filtering to the amount providing the best forecast quality with respect to, e.g., to the low-level wind field or orographic precipitation. Moreover, the previous approach of combining an SSO (subgrid-scale orography) parameterization with an orographic roughness length approach (i.e. enhanced roughness length in the presence of unresolved orography) has been replaced by using an (appropriately retuned) SSO scheme only, which avoids consistency issues between momentum and heat/moisture transfer in the turbulence scheme. The snow parameterization within the land-surface scheme has been revised in several aspects, providing a more realistic representation of snow albedo, snow density and heat conduction through the snow cover. In addition, a tile approach has been introduced in ICON that allows for a better representation of heterogeneous land cover and a partially snow-covered surface. Verification results from various regions of the earth confirm the advances in forecast skills related to these improvements.

MeteoSwiss develops the new numerical weather prediction (NWP) systems COSMO 1 and COSMO E to be operational in 2016. COSMO-1 will provide locally very detailed deterministic forecasts out to +24 hours using a mesh size of 1.1 km and will run 8 times a day, whereas the probabilistic COSMO E will provide an ensemble forecast out to +120 hours using a mesh size of 2.2 km twice a day. Both systems will run for a domain covering the Alpine area. The initial conditions (analyses) for these forecasts will be generated by a new ensemble data assimilation system based on a Local Ensemble Transform Kalman Filter (LETKF) providing analysis uncertainty as well. COSMO 1 as well as COSMO E already run experimentally on a daily basis and demonstrate their benefits mainly in complex topography where high-resolution information and small mesh-size are important for the simulation of atmospheric processes.

The talk will highlight findings from the development of COSMO 1 and COSMO E with emphasis on Alpine meteorology and the benefits of very-high resolution NWP. Results on improved representation of valley winds, idealised simulation of convection over topography, and stochastic physics perturbations will be shown. Additionally, first verification results of the experimental regular runs of the two systems will be presented, providing exemplary evidence that the use of ensemble forecasts (such as COSMO E) in combination to very-high resolution deterministic forecasts (such as COSMO 1) is advantageous for atmospheric processes with limited predictability. Last but not least, first experiences from running these new forecasting systems on a novel GPU-based high-performance computing architecture will also be discussed.

The bora wind is downslope wind which blows in a coastal part of Croatia. Due to gusts which are stronger and much often in a winter time, complex orography and land/sea transition in this area, the modelling can be challenging. Within the research with the NWP WRF model, various tests were made with implementation of the new, improved mixing length in MYJ PBL scheme. Turbulence parameterization in MYJ scheme is based on mixing length (ML) scale which is defined as a physical quantity describing the size of the most relevant eddies in a modeled turbulent flow. The new ML is uniformly valid in neutral and static stable airflows. Based on previous research with vertical diffusive schemes in numerical models, the correction of vertical diffusion has been implemented as well. During period from 01 Jan to 31 March of 2011 there have been 17 episodes of bora wind with typical duration longer than 10h. Using high horizontal and vertical resolution the model has been tested and evaluated with measurements from mast tower located in mountainous region in a hinterland on Pometeno brdo as well as with ground measurements from meteorological stations and soundings.

A new parameterization scheme to represent rain enhancement due to sub-grid orography via the seeder feeder mechanism will be described. Tests using idealized 2D simulations of the KiD (Kinematic Driver) model, developed at the Met Office, demonstrate that the scheme successfully increases the total rainfall over a low resolution hill towards that produces by a well resolved hill. Simplifying assumptions are made about the nature of the variation of the sub-grid surface height about the grid-box mean, enabling us to estimate the amount of orographic water missing from an NWP grid-box from the sub-grid orographic standard deviation and the grid-box mean RH. This sub-grid orographic water is added to the liquid water mixing ratio used to calculate the accretion rate, thereby enhancing the rain rate. The evaporation of rain is also reduced in the presence of sub-grid orographic cloud. The scheme can be extended to cold cases where snow is present instead of rain.

Currently the scheme is being coded into the Met Office Unified Model in order to test its performance for real cases of orographic rain enhancement. Results from these tests will be shown.

Mountainous basins throughout the world are impacted by prolonged episodes of boundary-layer air stagnation or persistent cold air pools (CAPs) during the winter season. These CAP episodes are often associated with poor air quality within urban basins. The detailed meteorological evolution of CAPs are generally inadequately resolved in mesoscale numerical weather prediction (NWP) forecast models, and even small model errors can have large impacts on temperature and air quality forecasts. In this study, we present a number of incremental improvements that have been applied to numerical modeling of these episodes within the Weather Research and Forecasting (WRF) model framework. Simple changes to the specification or parameterization of land use, snow cover, initial atmospheric and surface state, cloud cover, and turbulent mixing have resulted in improved CAP simulations. Several simulated case studies of CAP evolution and sensitivity to surface state during the Persistent Cold Air Pool Study (PCAPS) and Uinta Basin Winter Ozone Study (UBOS) will also be presented.

During the first HyMeX special observing period (Sep-Nov 2012), the LEANDRE II lidar on board the ATR 42 flew 22 missions aiming at collecting water vapor observations of the marine inflow upstream of heavy precipitation events forecast to hit the Spanish, French or Italian coasts. These lidar observations were assimilated using the 3D VAR assimilation system of the AROME numerical weather prediction mesoscale model. The assimilation was carried out for the period of 11 September-27 October by running a 3-hour forward intermittent assimilation cycle. In the presentation, the impact of the lidar observations will be assessed, first by comparing the new analyses with the control analyses, then by comparing the precipitation forecasts (obtained with and without the lidar observations) for the 25 rainy days included in the assimilation period. The impact of the assimilation of the water vapor observations is found to be positive up to the first 12 hours of the precipitation forecasts. Special emphasize will be put on the convective line of 24 Sept. for which the location and timing of the system appear very sensitive to the detailed distribution of the upstream moisture.

Accurate radar precipitation estimation and nowcasting for the Swiss Alps region have always been challenging due to the complexity of the terrain which may impede unobstructed observation. In 2013 MeteoSwiss launched “CombiPrecip”[1,2,3], a sophisticated raingauge-adjustment application of the radar precipitation estimation maps. The main purpose of this application is to produce rainfall estimation as close as possible to the ground-truth. CombiPrecip has been built on the basis of cutting-edge technology originated in the field of geostatistics. It uses co-kriging with external drift to merge spatiotemporal information from a small number of raingauge measurements with the rainfall field observed by the radar and produces a new improved precipitation map. Carefully constructed validation schemes which use a number of skill scores have shown that this improvement is both systematic and important. CombiPrecip is equipped with many novel modules that deal with problems such as convection control in locations and times of low raingauge representativeness, and adjustment over regions not covered by the raingauge network such as out of the country border.

In the field of nowcasting, efforts of MeteoSwiss are currently focusing on a generalization of the “MAPLE”[4], a Lagrangian-persistence-based application, originally developed in McGill University in Montreal. This generalization has been motivated by the fact that in the Alps the evolution of precipitation is strongly influenced by the mountains at several different scales and this influence depends on several aspects such as direction and strength of flow and air-mass stability[5]. Therefore our effort attempts to apply modern statistical learning techniques to numerous multiple-year-recorded variables in order to organize, stratify and eventually recognize repeatable localized weather patterns. Such information can be then coded efficiently and used in an operational scheme. This design is fully probabilistic, but the assignment of probabilities will be based on the history of patterns that have emerged through the statistical supervised learning process.

References

1. Sideris I.V., M. Gabella, R. Erdin and U. Germann, (2014) “Real-time radar-raingauge merging using spatiotemporal co-kriging with external drift in the alpine terrain of Switzerland”, Q. J. Roy. Meteor. Soc., 140 (680), 1097-1111.

2. Sideris I.V., M. Gabella, M. Sassi and U. Germann (2012) “Real-time spatiotemporal merging of radar and raingauge precipitation measurements in Switzerland”, Proceedings, 9th International Workshop on Precipitation in Urban Areas, St. Moritz, Switzerland.

3. Sideris I.V., M. Gabella, M. Sassi and U. Germann (2014) “The CombiPrecip experience: development and operation of a real-time radar-raingauge combination scheme in Switzerland”, Proceedings, 2014 International Symposium Weather Radar and Hydrology, Washington DC, USA.

4. Germann U. and I. Zawadzki (2002) "Scale-dependence of the predictability of precipitation from continental radar images. Part I: Description of the methodology." Monthly Weather Review 130.12, 2859-2873.

5. Panziera, L., et al. (2011) "NORA–Nowcasting of Orographic Rainfall by means of Analogues." Quarterly Journal of the Royal Meteorological Society 137.661, 2106-2123.

In the framework of the SATIN project, a nowcasting system has been developed and run over an Eastern Alpine domain at ZAMG. It is based on satellite derived precipitation estimates and applies similar algorithms as in the operational INCA nowcasting system, which is normally driven by radar and station observations. Different to the operational version, this satellite INCA system SAT-INCA can be run not only over the Alps, but also in more remote areas, where no radar data and only few conventional data are available. Furthermore, a 1km resolution version of the convection permitting numerical weather prediction model AROME has been tested over the same Alpine area for the first time. For this AROME version also highly resolved orographic data (SRTM), updated land usage data (ECOCLIMAP II) and different boundary conditions were tested. Several case studies in spring, summer and autumn 2014, capturing different synoptic situations like local thunderstorms, wide spread flooding, squall lines were investigated. An inter comparison and evaluation between the new systems, satellite driven INCA and AROME 1km against the operational systems, radar driven INCA , AROME 2.5km and the coarser resolved ALARO 4.8km NWP, has been conducted. Results show, that AROME 1km can better simulate small scale near ground temperature features in Alpine valleys than the coarser models while in flat terrain ALARO is closer to the INCA analysis. There is a very slight positive impact of new orography and new land use data on the precipitation forecast. The highly resolved AROME precipitation simulations are closer to the nowcasting system during the first hours of integration, than the coarser resolved models, but tend to overestimate it later especially over the Alpine Mountain Range. SAT-INCA has its strengths in the rapidly updated analysis which exhibit superior performance than the operationally available NWP fields at that time. Additionally, SAT-INCA nowcasting performs better than the tested NWP models (ALARO, AROME, AROME-1km) in the first hour. For longer integration times, NWP clearly outperforms the satellite based nowcasting.

In its standard version, the INCA analysis and nowcasting system uses motion vector fields for the nowcasting of precipitation. These fields are based on correlations between patterns in previous precipitation analyses and provide information for a kinematic extrapolation over the first few hours of lead time. By nature, these short range forecasts cannot account for changes in shape and intensity of precipitation structures which occur especially in convective situations.

Two new approaches have recently been implemented and tested at ZAMG, with the goal of improving the nowcasting quality.

1) Superimposing cell-tracking vectors over the motion vector fields: The A-TNT system is a radar-based, object-oriented approach for cell tracking that provides information on cell movement and size. The cell tracking vectors are superimposed over the conventional motion vectors wherever applicable and thus apply modifications to the original extrapolation.

2) Statistical advection: Based on a linear regression approach and calibration data from three convective seasons, predictors such as Lagrange persistence, radar data, convective parameters and others are used to derive modified nowcasting fields.

Objective evaluations using the SAL verification method and standard scores clearly showed improvements in the nowcasting range with both methods. The presentation will show preliminary results as well as further steps towards a refinement of the methodology and adaptations for operational use.

The increasing number of spatial verification methods raises the question of their properties and information content. Specific questions like - How does each method inform about forecast performance overall? Does the method inform about location errors? If so, how? Which methods yield identical information to each other? Which methods provide complementary information? – have already been addressed in a spatial forecast inter-comparison project running from 2008-2010 in a more statistical way and over flat terrain.

The second phase, called MesoVICT has been established to further explore the new methods for more realistic meteorological scenarios. MesoVICT has been launched as an official project of the Joint Working Group for Forecast Verification Research (JWGFVR) a working group of the WMO/WWRP. Its test cases include more variables in addition to precipitation, such as winds. Further, the cases represent interesting meteorological events that develop over time rather than single snapshots. The cases also include ensembles of forecasts as well as observations, and, as the name suggests, they are provided on the larger Alpine region, a region associated with complex terrain.

The aims of the project can be summarised as follows:

- To investigate the ability of existing or newly developed spatial verification methods to verify fields other than deterministic precipitation forecasts, e.g., wind forecasts and ensemble forecasts.

- To demonstrate the capability of spatial verification methods over complex terrain, and gain an understanding of the issues that arises from this more challenging situation.

- To encourage community participation in the development and improvement of spatial verification methods, especially for evaluating high resolution numerical forecasts.

- To provide a community test bed where common data sets are available, but also for the sharing of data and code to assist in developing and testing spatial verification methods.

The talk will give some information on the project structure of MesoVICT, data and models used and cases investigated.

Atmospheric processes associated with complex terrain include various phenomena on the meso- and microscale, which contribute significantly to the local weather in mountainous areas of the Earth. Phenomena such as the thermally induced valley wind system, cold pools, foehn winds and channelling of synoptic winds are well-known. However, they still pose a challenge to numerical weather prediction (NWP) models. Several factors, such as the horizontal and vertical grid resolution, input data, initialization, terrain representation and the parameterizations are important for a successful simulation of atmospheric phenomena in complex terrain. Meteo Swiss is currently testing a pre-operational setup of the NWP model COSMO (Consortium for small-scale modelling) on a horizontal resolution of 1.1 km for a domain over the main Alpine range.

In this contribution we assess COSMO's performance with respect to turbulent exchange processes in complex terrain. This is done by using data from the so-called “i-Box” in the Inn Valley, Austria. The “i-Box” consists of six core measurement sites that are located at representative locations in the Inn Valley, and two remote sensing systems (Lidar and HATPRO) in the city of Innsbruck. The long-term measurement set provides a data pool of high-resolution velocity variances, turbulence variables, radiation, soil moisture, and vertical profiles of temperature, humidity, and wind in the lower troposphere.

The model is assessed by the means of case studies of several weather situations representative for the Inn Valley, such as thermally forced flows, stable boundary layer formation, and foehn events. Two case study days will be presented in this contribution. With a process-oriented analysis, we investigate the spatial variability of the surface characteristics, the radiative forcing, the resulting near-surface turbulence structure, and the valley boundary layer development. A special focus is laid on the influence of the terrain representation on the performance of the model for several locations such as valley floor, mountain top, and slopes with different orientation and angle. The results show that the model's performance is indeed dependent on the location (valley/slope/mountaintop), but also on the time (day/night). This allows us to evaluate possible deficiencies of the model setup and to investigate whether the model's physics schemes (initially developed for horizontally homogeneous and flat surroundings) are suitable for truly complex terrain.

Precipitation datasets are used to evaluate climate models or to remove model output biases. Especially in mountainous regions, observed precipitation is error prone due to its high spatio-temporal variability and due to considerable measurement errors. Observational uncertainties are usually neglected in climate model evaluation or in bias correction studies. In this study we compare three datasets: the widely used pan-European gridded E-OBS dataset, the pan-European MESAN downscaled HIRLAM reanalysis (HMR), and a set of regional datasets that are based on high station densities and are partly corrected for undercatch. We show that the differences between these datasets can be larger than precipitation errors in the EURO-CORDEX regional climate model simulations. Differences between the observational datasets are especially large in regions where E-OBS and the HMR have low station densities, for low and high temperatures, high latitude, and mountainous regions (the Alps, Scandinavian Mountains, Carpathians, Pyrenees). These results highlight the strong need for using multiple observational datasets for model development, evaluation, and statistical post-processing as well as for impact research focusing on hydrology, agriculture, ecosystem management, or climate extremes. The magnitude of these so far not regarded observational uncertainties can severely affect climate change impact assessment and lead to misguided political decisions.

The knowledge of precipitation trends and variability is vital for many aspects of life and socio-economic sectors. However, confidence in precipitation trends is still limited and merits regular reassessment. Here, seasonal and annual homogenized precipitation series in Switzerland are investigated for the period 1901-2013 in terms of trends, interannual variability and the influence of large-scale European flow patterns. An objective spatial clustering is applied to 305 stations resulting in and 32 distinct precipitation regions. Moving window trend periods from 33 to 113 years length have been analysed on a seasonal and annual basis using Theil-Sen trend estimates and non-parametric Mann-Kendall trend tests (significance limit p<0.05).

Of 2720 analysed moving trend windows, 194 (7.1%) show a significantly positive trend, and 10 (0.4%) a significantly negative trend. Most of the significantly positive trends are found for long precipitation series (50+ years) for winter, autumn and annual series. 81 (72)% of the annual (winter) series in the 1901-2013 period show positive trends and for 34 (22)% of the regions the trends are significant. Significantly negative trends are only found in winter for some short time series in the most recent decades. Interannual variability is varying considerably regionally and seasonally. No clear long-term trends could be identified. The same conclusions hold for changes in moderate seasonal extremes.

The influence of large scale flow strongly depends on the season and region. The strongest link between large scale flow and Swiss precipitation variability is found for winter. With the variability of only four major patterns of sea level pressure, precipitation sums of individual years can be reconstructed with a mean relative error of about 15-25% for northern and 25-35% for southern Swiss regions. The most important pattern for northern Switzerland is a Euro-Atlantic blocking like pattern. For southern Switzerland, an Eastern Atlantic like pattern is dominant in winter and a Scandinavian like pattern in autumn. Composites of sea level pressure and near-surface wind anomalies for dry and wet seasons confirm the importance of atmospheric blocking in winter and autumn. In spring, a high pressure anomaly stretching from the mid-Atlantic to central Europe seems to be very indicative for dry years in northern Switzerland.

In the framework of the Interreg IV project 3P-Clim* trends of temperature and precipitation were examined for the region between Arlberg and the Dolomites. An emphasis was laid on the link between the changes, that have already been observed, and the projected changes.

Homogenised daily data series of 17 selected stations from start of the measurements until end of 2010 build the basis to analyse the already observed changes. The regional climate model (RCM) COSMO-CLM was used to downscale the ECHAM5 global scenario A1B (“realistic scenario”) to a spatial resolution of 10 km in a two-step approach. The climate forecast was utilised for the two thirty year periods 2026 – 2055 and 2071 – 2100.

Results:

It is shown, that the observed temperature trend is spatially quite consistent. The strongest warming in the annual mean temperature occurred in the period from mid 70-ies to about 2003. The warming is most evident in spring and summer, while in winter a more intermittent signal with embedded phases of cooling is seen. The projected temperature evolution basically conforms to the monitored development and predicts an increase of 1 – 2 (3 – 4) degrees for the period 2026 – 2055 (2071 – 2100) compared to the reference period 1981 – 2010. Examples for implications on derived climate indices like number of summer days, frost days etc. are given.

In annual mean precipitation generally high year-to-year variability with little to no significant trend is observed. Nevertheless, a slight tendency to an increase north of the Alpine ridge and a decrease south of it is found. This is in accordance to analysis of the Histalp data-set (Böhm 2008). The RCM predicts less precipitation in the whole area towards the end of the 21st century, with an even more pronounced decrease in the Veneto region. This reduction emerges mostly from the summer months where the decrease is most prominent.

Literature:

Böhm R, 2008: Harte und weiche Fakten zum Klimawandel – ein Überblick. In: Böhm R, Godina R, Nachtnebel HP, Pirker O, (Red.), Auswirkungen des Klimawandels auf die österreichische Wasserwirtschaft. Hrsg. vom BmLFUW und ÖWAV, Wien, 53-70.

* Past, Present and Perspective Climate of Tirol, Südtirol-Alto Adige and Veneto

Foehn is an ubiquitous and yet very local phenomenon in complex terrain. Foehn climatologies in the Alps have been compiled at only few locations and often in a way that one cannot directly compare them.

A new objective, probabilistic and mostly automatic method makes foehn diagnosis so efficient that it can be scaled up for many locations. We present our current standing on the way to an objective and comparable foehn climatology of the whole Alps. Frequencies and characteristics of foehn winds at stations at different locations on both sides of the main Alpine crest will be compared.

The geographical region from where a cyclone is propagating into Europe appears to play an important role in generating certain weather extremes. Some of the most devastating European floods have been associated with type Vb cyclones, as in August 2002 or with a congeneric type in June 2013 for example. The aim of this work is to assess different propagation paths of atmospheric cyclones in terms of a systematic relation to large scale extreme precipitation events over Central Europe (CE) and examine future changes of precipitation extremes.

Part A: The relevance of cyclone tracks types for heavy large scale precipitation events and the change of cyclone track type frequency under future climate conditions:

Cyclone tracks have been determined form the analysis of vorticity- and gradient fields, derived from sea level pressure and geopotential height. Resulting tracks are then classified into nine types by a new stream-based approach, based on the geographic regions from where cyclones enter CE. The different types are evaluated in terms of a systematic connection to large scale precipitation events afterwards.

Results show very clear differences between different cyclone track-types in terms of their climatological characteristics, as well as in their relevance for large scale precipitation amounts. The exceedance probability for a heavy precipitation event is very different between different sub-domains and between different track types, although the study region covers a rather small part of CE. The most important track types in this respect are Vb, X-N, X-S in the southeastern, as well as ATL and TRZ in the northwestern regions.

Global Climate Model simulations (Echam6, Echam5, EC-Earth) show a good agreement between the observed and modeled track type frequency. Under future climate conditions all models show a clear decrease in the number of cyclone tracks passing Central Europe, especially in summer. Signals for individual track types are not that clear, as differences between models are usually larger than differences arising from different GHG-concentration scenarios, due to the large internal climate variability.

By using the twice-daily atmospheric observation data from 1998 to 2012, station rainfall data, Tropical Rainfall Measure Mission (TRMM) data, as well as the plateau vortex and shear line year book, characteristics of the sustained departure plateau vortexes (SDPVs) are analyzed. Some new useful observational facts and understanding are obtained about the SDPV activities. The following results are obtained. (1) The active period of SDPVs is from June to August, most in July, unlike that of the unsustained departure plateau vortexes (UDPVs), which have same occurrence frequencies in the three summer months. (2) The SDPVs, generated mainly in the Qumalai neighborhood and situated in a sheared surrounding, move eastward or northeastward, while the UDPVs are mainly led by the upper-level trough, and move eastward or southeastward. (3) The SDPVs influence wide areas of China, even far to the Korean Peninsula, Japan, and Vietnam. (4) The SDPVs change their intensities and properties on the way to the east. Most of them become stronger and produce downpour or sustained regional rainstorms to the south of Yellow River. (5) The longer the SDPV sustains, the more baroclinity it has. (6) When an SDPV moves into the sea, its central pressure descends and rainfall increases in all probability. (7) An SDPV might spin over the bend of the Yellow River when there exists a tropical cyclone in the East China Sea. It could also move oppositely to a landed tropical low pressure originated from the sea to the east of Taiwan or from the South China Sea.

Moist convection is a fundamental process in our climate system, but is usually approximated in climate models by using semi-empirical parameterizations. These approximations introduce significant uncertainties and biases, and there is thus a general thrust towards higher resolution and the explicit representation of moist convection. This approach had been pioneered in research-style simulations many decades ago, and has been used in operational numerical weather prediction for a few years. For climate applications, convection-resolving simulations are still very expensive, but they are increasingly becoming feasible and attractive.

Here we present recent results pertaining to the development and exploitation of convection-resolving atmospheric models for climate scenario applications. We discuss the potential and challenges of the approach, highlight validation using decade-long simulations, and present recent results regarding the application to climate change projections. Most examples will be use a 2.2 km horizontal resolution and cover an extended Alpine region from Northern Italy to Northern Germany.

Consideration will also be given to future challenges, as related to the use of filtered governing equations, the representation of turbulence, and the use of emerging heterogeneous many-core hardware architectures. Some preliminary results will be presented of simulations using an extended computational domain covering Europe.

Extreme value analysis provides a useful tool for the estimation of the probability of unusually large precipitation events that can cause devastating floods. Climate modeling studies, as well as theory and observations suggest that such extreme precipitation events will intensify with warming. Current projections of extreme precipitation are based on conventional global and regional climate models, which due to the coarse resolution need to parameterize convective precipitation. The use of convection parameterization leads to a poor representation of extreme precipitation especially on the sub-daily time scale.

Here we present an analysis of extreme precipitation events in convection-resolving climate simulations. The simulations are performed with the COSMO-CLM model at a convection-resolving resolution of 2.2 km across an extended Alpine domain. Ten-year long control and scenario simulations (1991-2000 and 2081-2090) are conducted, driven by a CMIP5 coupled climate model (MPI-ESM-LR) under an RCP8.5 greenhouse gas scenario. Generalized extreme value theory is applied to address projections of daily and hourly extreme precipitation events in the winter and summer seasons. We present a detailed intercomparison of convection-resolving model against a conventional (12 km grid spacing) climate model in terms of model performance and difference between climate change signals.

Global-warming-induced changes in extreme orographic precipitation are investigated using a hierarchy of models: a global climate model, a limited-area weather forecast model, and a linear mountain-wave model. We consider the changes produced by a doubling of CO2 in precipitation over idealized north-south mid-latitude mountain barriers at the western margins of otherwise flat continents.

The intensities of the extreme events on the western slopes increase by about 4%/K of surface warming, close to the ``thermodynamic" sensitivity of vertically integrated condensation in those events due to temperature variations when vertical motions stay constant. In contrast, the intensities of extreme events on the eastern mountain slopes increase at about 6%/K. This higher sensitivity is due to enhanced ascent during the eastern-slope events, which can be explained in terms of linear mountain-wave theory as arising from global-warming-induced changes in the upper tropospheric static stability and the tropopause level. Similar changes to these two parameters also occur for the western-slope events, but the cross-mountain flow is much stronger in those events, and as a consequence, linear theory predicts no increase in the western-slope vertical velocities.

Extreme western-slope events tend to occur in winter, whereas those on the eastern side are most common in summer. Our idealized climate-model simulations also suggest that changes in extreme precipitation over the mountains will be smaller than that over the oceans or over flat land.

Mid-latitude mountain regions are particularly sensitive to climate change because of an active snow-albedo feedback (SAF). Here, the SAF is characterized and quantified over the complex terrain of the Colorado headwaters (HW) region of North America using high-resolution WRF regional climate model simulations. A pair of 7-year control and pseudo global warming (PGW) simulations are used to study the regional climate response to a large-scale climate change.

Warming is strongly enhanced in regions of snow loss by as much as 5°C. Linear feedback analysis is used to quantify the strength of the SAF within the HW region. The strength of the SAF reaches a maximum value during April when snow loss coincides with strong incoming solar radiation. Simulations using 4 km and 12 km horizontal grid spacing show good agreement in the strength and timing of the SAF, whereas a 36km simulation shows discrepancies that are tied to differences in snow accumulation and ablation caused by smoother terrain.

Energy budget analysis shows that transport by atmospheric circulations act as a negative feedback to regional warming, damping the effects of the SAF. On the mesoscale, the SAF non-locally enhances warming in locations with no snow, and enhances snowmelt in locations that do not experience snow cover change. The methods presented here can be used generally to quantify the role of the SAF in simulated regional climate change, illuminating the causes of differences in climate warming between models and regions.

Beginning with the work of Blanford in 1884, the Tibetan Plateau has been recognized as having a profound impact on the intensity of the Indian Summer Monsoon. However our ability to fully characterize the changing nature of the plateau’s impact on the monsoon has been limited by the lack of observations of the hydroclimate of this remote and data sparse region. The recent availability of atmospheric reanalyses that assimilate only surface pressure data now allows the possibility to extend our knowledge of this relationship back in time. In this presentation, we will make use of the 20th Century Reanalysis from the ECMWF (ERA20C) along with the APHRODITE high-resolution gridded precipitation dataset and an ice core extracted from a high elevation site on the Dasuopu Glacier on Shishapangma, an 8000 meter high peak in Tibet situated just north of the Himalaya, to examine trends and multi-decadal variability in the hydroclimate of Northern India and Tibet since 1850.

In particular we show that over the 20th Century, the ERA20C indicates that there has been a statistically significant decrease in annual mean snowfall over the southern Tibetan Plateau and a concomitant increase in annual mean rainfall over Northern India that extends northwards into the Himalaya. Both of these changes are shown to be associated with the enhanced warming that has occurred over the plateau since the start of the 20th Century. These results are consistent with trends in rainfall and snowfall as recorded in the APHRODITE dataset over the last 50 years. They are also consistent with a long-term trend towards lower snow accumulation in the Dasuopu ice core that began in the middle of the 19th Century. We also use an EOF analysis as well as spatial correlation analysis to examine multi-decadal variability in rainfall and snowfall in the region of interest and the changing nature of the relationship between the two.

Permafrost is a widespread phenomenon in the European Alps and is characterised by temperatures only a few degrees below zero and is therefore particularly sensitive to projected climate changes in the 21st century. To evaluate the sensitivity of mountain permafrost to climatic changes and to assess its future evolution, not only climatic variables such as air temperature, radiation and timing and duration of snow cover have to be considered, but also subsurface characteristics such as ground temperature, ice content, porosity or hydraulic properties. Permafrost monitoring in the Swiss Alps started only 1-2 decades ago, but currently comprises a large set of meteorological, geophysical, kinematic and ground thermal parameters at a large variety of field sites. The large Swiss project cluster TEMPS (The evolution of mountain permafrost in Switzerland), funded by the Swiss National Science Foundation, integrated and analysed this data set for the first time and combined the observations with long-term model simulations using a dynamic process-oriented permafrost model (COUP-model). In combination with results from Regional Climate Model ensembles, TEMPS aimed to create plausible evolution scenarios of mountain permafrost at specific sites and investigated the interactions between atmosphere and permafrost focusing on the evolution of ground temperature, ice content and related degradation and creep processes.

In this contribution we will show the main results of the project by focusing on several Swiss monitoring stations with permafrost temperatures close to the melting point. Long-term simulations with the coupled heat and mass transfer model COUP forced by downscaled GCM/RCM chains from the ENSEMBLES project suggest increasing ground temperatures and permafrost thaw until the end of the century. The projected air temperature increase leads also to a corresponding reduction in snow cover (= decrease of surface insulation leading to a potential cooling effect), however, the latter does not offset the general warming trend of permafrost temperatures in the model simulations. The high variability of surface and subsurface materials in the permafrost regions of the European Alps will strongly modulate any general trend which might already be visible within the coming decades. Furthermore, we will present technical improvements regarding (a) strategies for downscaling and debiasing RCM output data for permafrost analysis on the station scale at high altitudes and (b) new geophysical and kinematic observation techniques for high-mountain areas.

The Vatnajökull ice cap at the southeast coast of Iceland has a large effect on the spatial distribution of precipitation in Iceland. This effect is quantified based on results from two sets of high-resolution atmospheric simulations. A control simulation employs the current land height and glacial cover while in a sensitivity run the ice caps have been removed and the height of the model surface is based on the bedrock topography of the glaciers. The simulations are done at 8 and 2 km horizontal resolution and are forced with the Interim re-analysis of the ECMWF for two consecutive years 2004-2006.

The simulations indicate that in the absence of the Vatnajökull ice cap, up to 25% decrease in annual precipitation may be expected in large regions covered by the ice cap. The overall decrease for the whole of Vatnajökull may be close to 15% when the glacial cover is removed. The reduced topography height leads to greater spillover, i.e. a larger part of the precipitation associated with lows and fronts approaching from the south and the east reaches the northern and western ice margins. There are, however, only minor changes further into the lee of the ice cap. The results of this study are not only of relevance in general studies pertaining to atmospheric flow and complex terrain, but also in e.g. planning of hydropower availability and harnessing in a warming climate characterized by retreating glaciers.

Climate models provide a rich set of information for use in studying the climate system and its possible evolution; however, the size of the grid cells in climate models makes it impossible for them to explicitly represent important alpine processes. As a result, making use of climate change information in alpine environments requires some degree of post-processing of the output. Post-processing options are often viewed as a dichotomy between statistical and dynamical methods, but in reality, there exists a continuum of methods that have developed in recent decades. These methods have traditionally ranged from relatively simple bias corrections designed to mimic statistics of current climate, to complex statistical fits of circulation patterns designed to preserve more physical meaning, and finally to highly complex regional climate models (RCMs) based on first principles. However, an end-to-end physically based method has eluded the applications community due to the tremendous computational cost associated with RCMs. Recently, we have developed the Intermediate Complexity Atmospheric Research model (ICAR) to fill the gap between advanced statistical methods and full RCMs. ICAR shares much in common with a state-of-the-art RCM, but requires less than 1% of the computational cost. Here we compare the climate change signal in the Colorado Rocky Mountains as predicted from methods spanning the downscaling complexity continuum. Finally we discuss the implications of the range of outputs from different methods; how should one make use of this potential glut of information and methods? How can we know which change signals may be more reliable?

Forecasting temperature variation at small scales in complex terrain is important for predicting localised hazards such as freezing temperatures on sections of road, cold damage to crops, and future changes to complex terrain microclimates. Cold air pools which form in valley bottoms on calm, clear nights drive exaggerated low temperature extremes compared to flat terrain. Many valleys in which cold air pools form, however, are unresolved by current operational numerical models, so that predictions of temperature must be derived by using post-processing techniques to downscale the coarse operational model forecast. A scheme used at the Met Office is designed to account for the influence of sheltering on cold air pooling in sub-grid valleys, and can be applied to create fine scale maps of forecast screen temperature. An improved version of this scheme is described, including a component which targets improved upland temperature prediction via a treatment of the interaction between the hilltop boundary layer and the horizontally adjacent atmosphere above the valley floor.

Statistical post-processing is often used to downscale a global weather prediction model to a particular location, especially in complex terrain like the Alps. When expanded to several stations, the post-processing has to be repeated at every station individually, typically ignoring the available spatial information and thus potentially leading to spatially incoherent forecasts. To maintain spatial coherence, the post-processing is applied to all stations simultaneously. Therefore, site-specific characteristics are eliminated by subtracting a spatial climatology from measurements and numerical forecasts. This results in anomalies on which simple linear regression can be performed for multiple stations simultaneously. Furthermore, since these spatial climatologies are made with generalized additive models for the whole region at once forecasts at any location are possible even where no measurements are available.

The Analog Ensemble is statistical post-processing technique to generate probabilistic forecasts by searching similar past numerical weather predictions (i.e., analogs) across several variables (i.e., predictors) to the current prediction. The measurements corresponding to the best analogs form the analog ensemble (AnEn) with which the probability distribution of the future state of the atmosphere can be estimated. This study explores the application of AnEn for probabilistic short- or medium-range forecasts in complex terrain over Croatia.

The AnEn is generated with the Aire Limitée Adaptation dynamique Développement InterNational model (ALADIN) run over two nested domains with 8 and 2 km horizontal resolution, respectively. It is tested at several climatologically different locations across Croatia for point-based wind speed predictions at 10 m and 80 m height. Results are verified and compared to the ALADIN model to address the following question: how does AnEn performs at locations in complex terrain over Croatia? The analysis focuses on a group of stations with downslope windstorms occurrences such as bora wind.

The verification procedure includes several metrics (e.g., Brier skill score, ROC skill score, reliability and dispersion diagrams) to optimize the AnEn configuration and to test the probabilistic prediction performances. Different predictors such as wind speed, direction, temperature, Richardson number and Scorer parameter are examined. Skill of AnEn predictions are compared with forecasts generated via logistic regression (LR). This study shows that the AnEn adapts well to different terrain and height. It provides accurate predictions while reliably quantifying their uncertainty and showing satisfactory spread. The AnEn performance is equal or superior than LR, especially for group of stations that are climatologically prone to strong winds. These results encourage the use of AnEn in an operational environment at meteorological station locations, as well as at wind farms.

Ensemble prediction systems (EPS) attempt to reflect the uncertainty of weather forecasts. To downscale EPS forecasts to point locations, and to correct for systematic errors and missing representativity, they are often statistically post-processed. Precipitation is a challenging quantity, because its distribution is non-Gaussian, bounded at zero, and contains a large fraction of zeros for short accumulation periods (e.g. 3, 6, 24-hourly sums). We present a new statistical EPS post-processing method, which masters these challenges. The idea is to employ a zero-censored Gaussian distribution to appropriately capture the point mass at zero precipitation amounts. Mean and scale of the underlying Gaussian distribution are then linked to the ensemble mean and scale, and estimated by maximum likelihood. Results are presented for the Alto Adige region in the North of Italy for short-range weather forecasts.