[University home]

Centre for Atmospheric Science

COPS Data Analysis and Model Studies

Data analysis and modelling work focusses on addressing the five components of the COPS science problem.


1: Orographic flow leading to development of convective cells

Our ability to model thermally driven flows in complex hill/valley systems is poor because of the lack of good observations of all the significant components of the surface heat and moisture balances and the frequent inadequacy of the understanding and representation of these processes in larger scale models. Significant processes to consider include full representation of the surface energy balance (including slope/shading effects), good representation of the boundary-layer turbulence, the influence of the overlying atmospheric wind and thermodynamic structure.

In COPS these shortcomings are addressed with a detailed and comprehensive set of measurements at multiple sites around the region. The relative importance of these processes and the complex interactions between the surface fluxes, the orography and the thermally-driven flow will be addressed through the synthesis of the experimental data in conjunction with numerical simulations using two complementary models — BLASIUS and the UM.

• BLASIUS is a non-hydrostatic model designed for idealised studies of boundary layer processes over hills. The original model, from the UK Met Office, has recently been developed at Leeds to include the addition of a bulk microphysical scheme, and a more sophisticated surface energy balance scheme. In addition, a new pressure solver has been developed at the Met Office to be able to deal with steep slopes. The model will be used for high resolution process studies using idealised orography and a range of atmospheric profiles.

• The UM is the UK Met Office operational forecast model. As such, it is more suitable for performing real case studies using real orography and operational UM analyses. Current versions include improved microphysics and prognostic rain.

A series of BLASIUS runs will be conducted over an idealised mountain/valley orography for a range of idealised atmospheric profiles to determine flow patterns over and around the orography and will quantify the fluxes of moisture and heat from the valleys into the sub-cloud layer. In particular the runs will investigate what boundary layer parameters determine whether the anabatic flow penetrates the free troposphere and where the convergence lines / zones occurs. These will indicate likely spots for convective initiation. Simulations of the diurnal cycle over the same idealised orography will be used to investigate the importance of soil moisture, vegetation, surface radiation (modified by slope, aspect and shading) and initial atmospheric structure on the flow within the mountain / valley system. Flow patterns will be qualitatively compared with observations of temperature, humidity and winds made using aircraft and remote sensing instruments. Such simulations will demonstrate which factors are important to include in an operational surface flux scheme and where the greatest sensitivity to initial conditions may be.

UM runs with real orography, but using idealised initial profiles, will be used to investigate how the small scale details of a more complex orography will affect the local flow. Real case studies will also be run using the UM. These will be at a high resolution of 250 m (compared to a current operational resolution of 4 km over the UK). Using the extensive observations available in this case we will be able to quantify any forecast errors and investigate whether they are related to the dynamics / thermodynamics or the microphysics. We will also be able to test whether modifications made to the surface flux schemes in light of the more idealised experiments improve model performance and examine how well the new microphysics scheme, applied at these high resolutuons compares with observations. In addition to using the model runs to evaluate and improve current surface flux schemes, simulations will also investigate the transport processes within the boundary layer. In particular they will be used to investigate the pathways for heat and moisture and aerosols to reach regions of convective activity.


2. Interpretation of Ground-based aerosol measurements at Hornisgrinde

The focus of this part of the project is to use the results of the ground based measurements to make a detailed characterisation of the aerosol that can be linked to the aircraft measurements, which whilst comprehensive cannot characterize the aerosol completely. This information will be used as input in the cloud modelling studies. This characterisation will include an assessment of giant and ultra-giant aerosols that may be crucial in the initiation of warm rain.

Ground-based field measurements of the physical and chemical properties of aerosol particles were made throughout the intensive period at the Hornisgrinde site as described above. This hill top location is an ideal site to represent the inflow to the convection developing in the region. The data collected will be synthesised in this work and a detailed hygroscopic and cloud droplet activation closure study will be performed with a range of models of varying complexity.

Recent studies indicate that in sub-saturated environments the water content of particulate is dictated almost entirely by the inorganic fraction and the organic fraction plays only a minor role. This study will further test this hypothesis using a combination of detailed aerosol measurements and a range of models.

To date, there are no adequate closure studies of CCN activity from basic physical and chemical measurements of the aerosol that have been conducted under supersaturations that are representative of real clouds. Further, recent studies have proposed that particle number dominates activation and composition is not important, however these were carried out at high supersaturations which are not representative of the real atmosphere. Under such conditions it is no surprise that particle number is the most important parameter. This work aims to probe real aerosol particles and real clouds under real conditions and hence, aims to probe the extent to which a mixed organic-inorganic aerosol departs from ideal behaviour and to assess our ability to predict such a departure. Heterogeneous ice processes will also be important in the cumulus clouds being studied in COPS. It is therefore important that the population of ice nuclei is studied in some detail. The approach will be to characterize the IN in the atmosphere and use this information to firstly determine if there is a link between different cloud behaviour and the aerosol composition and secondly as input to the CRMs. Soot, organic and biological material will be observed at high time resolution during COPS and offline samples will be analysed for a range of particle types, including dust.


3: Convective transport of aerosols out of the boundary layer

It is possible that anabatic flows will reach directly into the free troposphere and/or that orographically modified and enhanced boundary-layer turbulence (including gravity-wave induced turbulence) will provide mixing between the boundary layer and free tropospheric air. This work aims to: determine the role of boundary-layer convective elements (thermals, detrainment mechanisms, clouds) in mixing aerosols into the free troposphere; characterise lids, propagating waves along lids, and interaction of convection with the lids; determine the role of surface conditions on BL turbulence and representation in UM; determine the factors affecting the ability of the Met Office UM to reproduce the observed transport of aerosols into orographically-locked convection.

The development of thermals into small cumulus and cumulus congestus clouds and thus the vertical transport of aerosols, depends on the properties of the air that feeds the clouds. There is evidence from recent analysis of data gathered in the Convective Storm Initiation Project that thermal updrafts are typically narrow, warm and moist. In fact it is likely that air from near the surface survives in the core of thermals. However, regions of enhanced moisture are broader than that of the enhanced updraft speed which is probably a result of detrainment. The cores of boundary-layer thermal should therefore also contain enhanced concentrations of aerosols. This work will use measurements of aerosol size distributions and number from aircraft flights made in the BL to see if there are enhanced concentrations of aerosols within the cores of thermals. Convection is usually initially inhibited by one or more layers of stable dry air (lids). The clouds have to penetrate past the lids in order to develop from small cumulus into cumulus congestus. It is possible that this occurs as a result of increased surface heating during the morning. It is also possible that the interaction of the cloud with the lid causes the lid to weaken. Detrainment of moisture and hence (processed) aerosols occur at the location of lids. So it is important to characterise the lids to understand the development of convection and the resulting detrained aerosols. We will focus on the following questions : How far do convective turrets need to penetrate to induce irreversible mixing of boundary layer aerosols into the free troposphere? Is most of the transport due to a few large events or do smaller, but more frequent (and shallower) events contribute? How does orography influence this transport? Do breaking gravity waves make a contribution at all? What influence does the synoptic scale exert?

Synthesis of data gathered from the ground-based radiosonde stations, Salford Doppler lidar and German lidars, FGAM wind profiler, the BAe 146 and from the Meteosat Second Generation (MSG) satellite will be combined with the results of data synthesis and detailed modelling. Specifically, the BLASIUS model will be used at about 50-m resolution to examine the properties of these boundary layer structures, small cumulus clouds and the interaction with lids. Sensitivity studies will be performed on the resolution. Properties of aerosols are being measured at the ground-based stations. The BAe 146 will make measurements of aerosol composition and size, CCN and IN, and wind velocity, temperature and humidity during long legs below cloud base. Penetrations will also be made through cloud at all levels beginning just above cloud base. The link will be provided by the German DO-128 (Project Partner, Dr. Ulrich Corsemeier) which will fly in the middle of the boundary layer and just below the top of the boundary layer. The UK Doppler lidar and German lidars and the BAe 146 and German DO-128 will provide information on the properties of the thermals (scale length, frequency, depth and maximum altitude) and the aerosols contained within them. The lidars will enable us to examine better the spatial seperation and development of thermals and their relationship to the surface heteorogeneity. The distribution of lids will be measured primarily with the FGAM wind profiler, radiosondes and the aerosol lidars at a number of the COPS sites. The wind profiler will also measure the growth of the turbulent boundary layer during the day and record the passage of gravity waves. The lidars will also reveal whether there are elevated layers of aerosols within or between lids. Satellite IR and visible images will be used to determine where convective cloud is breaking through the lids, allowing their position to be related to the model simulations. Synoptic-scale influences (e.g. from descending dry layers) will be included in this analysis from radiosonde profiles and forecast model analyses. Unravelling the effect of aerosol on cloud development from the effect of changes in atmospheric structure requires that the diurnal evolution of convection be followed for a range of lid morphologies as well as aerosol amounts.


4: Microphysics and dynamics of convective clouds

In addition to the details of the orographic flows it is necessary to understand the properties and variability of aerosol particles being drawn into the convective clouds in order that the microphysics and dynamics can be understood, and to establish the influence of aerosols on the vigour and depth of the clouds and the intensity of the precipitation. The latent heat of freezing for example can be sufficient to allow clouds to grow much deeper.

In this work we aim to understand the microphysics and dynamics of the convective clouds through the growth stage. The mature cloud, which cannot be penetrated by the aircraft, will be studied by the German groups using the various radars. It is likely that most of the precipitation in this region forms through the ice phase. Even so, the production of raindrops through collision and coalescence may be important. Raindrops can speed up the glaciation process by becoming instant rimers thus allowing the Hallet-Mossop process to commence earlier. Understanding the properties of aerosols is crucial for both the formation of cloud drops that play the crucial role in the growth of precipitation particles and the direct production of ice particle by primary nucleation. The key questions this work aims to address are: How does the aerosol entering the cloud affect the development of the cloud and the precipitation? Are raindrops produced by collision and coalescence and, if so, what role do they play in the production of precipitation via the ice phase and the glaciation of the cloud? What is the relative roles of ice nuclei and secondary ice production processes?

These questions will be addressed through data synthesis in conjunction with model runs. The principal source of data will be from instruments on the FAAM BAe 146. It will be equipped with the aerosol mass spectrometer (AMS), CCN probe, VACC (volatility) and standard cloud microphysics instruments (PCASP, Cloudscope, Fast FSSP, 2DC, 2DP, Cloud Particle Imager and Small Ice Detector) in order to study the growth of cloud droplets, the formation and growth of ice particles and precipitation particles in the context of the detailed dynamics of the orographic convective clouds. Radar data will also be used to map the 3D structure of precipitation within the clouds.

Three different models will be used to study the cloud microphysical processes, the Met Office Cloud Resolving Model (CRM), the Explicit Microphysics Model (EMM), and the new Microphysics Cloud Resolving Model (MCRM). The CRM and EMMwill be run together as the dynamics from the CRM are required as input for the EMM. The combination will be used to examine the detailed microphysics and dynamics of the convective clouds. The MCRM will be used to specifically examine to the influence of aerosols on the intensity of precipitation and the dynamics of the clouds.


5: Forecasting Convective Precipitation

The predictability of convection over complex terrain is a difficult and important question. It depends upon the complex interactions between: 1. Mesoscale uncertainty, probably dominated by uncertainty in upper level potential vorticity (PV) and lower level wet-bulb potential temperature (qw); 2. Mesoscale response to resolved surface forcing; 3. Model uncertainties resulting from uncertainty in unresolved processes (especially boundary-layer flows) and surface fluxes; and 4. Model uncertainties resulting from the lack of case-specific information such as CCN and IN spectra, as well as inadequacies in representation of the processes which would use this information. These may have a significant impact on first-generation cloud evolution (if not initiation) and hence initiation of subsequent generations via e.g. downdraft processes. The objectives of this work are: To quantify the predictability of severe convective storms encountered during COPS; To evaluate the role of different sources of uncertainty that reduce predictability, including (i) mesoscale dynamics setting the environmental conditions, (ii) the boundary layer response to resolved surface forcing, and (iii) the role of microphysics in generating downdrafts and subsequent convection; To evaluate how the data synthesized from the COPS field measurements enhance predictability of the UM, and hence to establish the degree to which initialization and boundary data control predictability; To evaluate the predictability of the UM forecasts of convective storms over a nine month period.

This work package will focus on predictability within the UM at O(1 km) resolution. An ensemble approach will be taken to the simulation of selected COPS events. It will be essential to quantify the magnitude of different sources of uncertainty. Mesoscale uncertainty will be derived using initial and boundary data from the Met Office Global and Regional Ensemble Prediction System (MOGREPS). Uncertainty in small-scale surface forcing will be simulated by varying the surface fluxes; the amplitude and scale of this variability will be informed by results from COPS field data and idealised results. Uncertainty in microphysics will also be informed by results from COPS measurements. The US ARM mobile facility will be operating continuously at supersiteM from 1 October to 31 December 2007 providing high resolution (1 minute/60m) profiles. US scientists have agreed to incorporate the ’Cloud- NET’ analysis developed at Reading within ARM; variables such as cloud fraction liquid and ice water content are retrieved from the observed profiles and then compared with the values of these prognostic variables held within operational forecasting models. Monthly skill scores for model performance will be derived for a range of forecast lead times so that the impact of any proposed changes to cloud parameterisation schemes in the UM can be rapidly and objectively quantified and the degree of predictability established. This work can be seen as an integration of results from the other areas of work within COPS and forms the bedrock to the design of a practical ensemble based forecast system. It is anticipated that the results (and methodology) will be of direct value in determining the emphasis to be placed on different processes in a practical ensemble system.