Search

In numerical weather prediction (NWP), observations and models are quantitatively compared for the purposes of data assimilation and forecast verification. The spatial and temporal scales represented by the observation and model may differ and this results in a scale mis‐match error which may be biased and correlated. The aim of this paper is to investigate the structure of representation error in convection‐permitting NWP models for four meteorological variables: temperature, specific humidity, zonal and meridional wind. We use high resolution data from the experimental Met Office London Model (approximately 300 m grid‐length) to simulate perfect observations and lower resolution model data. The scale mis‐match error and its bias, variance and correlation are calculated from the perfect observation and low‐resolution model equivalents. Our new results show that the scale mis‐match bias is significant in the boundary layer for temperature and specific humidity, whereas the variance is significant in the boundary layer for all analysed variables. Furthermore, they are shown to be related to the mismatch in the high‐ and low‐resolution orography. Contrary to previous studies using low‐resolution, (km‐scale) data, horizontal correlations are shown to be insignificant. However, all variables exhibit considerable vertical representation error correlation throughout the boundary layer; for temperature a significant positive vertical correlation persists for all model levels in the troposphere. Our results suggest that significant biases and vertical correlations exist that should be accounted for to give maximum observation impact in data assimilation and for fairness in model verification and validation.

Operational forecasting centres increasingly rely on convection-permitting NWP simulations to assist in their forecasting of convective events. The evaluation of upgrades in the underlying NWP modeling system normally happens through routine verification using traditional metrics on two-dimensional fields, such as gridded rainfall data. Object- and process-based evaluation can identify specific physical mechanisms for model improvement, but such evaluation procedures normally require targeted and expensive field campaigns. Here, we explore the potential use of the UK operational radar network observations and its derived 3D composite product for evaluating the representation of convective storms in the Met Office Unified Model. A comparison of the 1 km x 1 km x 0.5 km 3D radar composites against observations made with the research-grade radar at Chilbolton in the southern UK indicates that the 3D radar composite data can reliably be used to evaluate the morphology of convective storms. The 3D radar composite data are subsequently used to evaluate the development of convective storms in the Met Office Unified Model. Such analysis was heretofore unavailable due to a lack of high-frequency three-dimensional radar data. The operational nature of the UK radar data makes these 3D composites a valuable resource for future studies of the initiation, growth, development, and organisation of convective storms over the UK.

With a number of operational centres looking forward to the possibilities of “city scale” NWP and climate modelling it is important to understand the behaviour of order 100m models over cities. A key issue is how to handle the representation of partially resolved turbulence in these models. In this paper we compare the representation of a clear convective boundary layer case in London in 100m and 50m grid-length versions of the Unified Model (MetUM) with observations. Comparison of Doppler lidar observations of the vertical velocity shows that convective overturning in the boundary layer is broadly well represented in terms of its depth and magnitude. The role of model resolution was investigated by comparing a 50m grid-length model with the 100m one. It is found that, although going to 50m grid-length does not greatly change many of the bulk properties (mixing height, heat flux profiles, etc.) the spatial structure of the overturning is significantly different. This is confirmed with spectral analysis which shows that the 50m model resolves significantly more of the energetic eddies, and a length scale analysis that shows the 50m and 100m models produce convective structures 2-3 times larger than observed. We conclude that, for the MetUM, model grid-lengths of order 100m may well be sufficient for predicting many bulk and statistical properties of convective boundary layers however the details of the spatial structures around convective overturning in these situations are likely to be still under-resolved. Spin up artefacts emanating from the inflow boundary of the model are investigated by comparing with a smaller 100m grid-length domain which is more dominated by such effects. These manifest themselves as along wind boundary layer rolls which produce a less realistic comparison with the lidar observations. A stability analysis is presented in order to better understand the formation of these rolls.

Convective initiation is a challenge for convection‐permitting models due to its sensitivity to sub‐km processes. We evaluate the representation of convective storms and their initiation over South Africa during four summer months in Met Office Unified Model simulations at 1.5‐km horizontal grid length. Storm size distributions from the model compare well against radar observations, but rainfall in the model is predominantly produced by large storms (50 km in diameter or larger) in the evening, whereas radar observations show most rainfall occurs throughout the afternoon, from storms 10‐50 km in diameter. In all months, modelled maximum number of storm initiations occurs at least 2 hours prior to the radar‐observed maximum. However, the diurnal cycle of rainfall compares well between model and observations, suggesting the numerous storm initiations in the simulations do not produce much rainfall. Modelled storms are generally less intense than in the radar observations, especially in early summer. In February, when tropical influences dominate, the simulated storms are of similar intensity to observed storms. Simulated storms tend to reach their peak intensity in the first 15 minutes after initiation, then gradually become less intense as they grow. In radar observations, storms reach their peak intensity 15‐30 minutes into their life cycle, stay intense as they grow larger, then gradually weaken after they have reached their maximum diameter. Two November case studies of severe convection are analysed in detail. Higher resolution grid length initiates convection slightly earlier (300 m cf. 1.5 km) with the same science settings. Two 1.5‐km simulations that apply more sub‐grid mixing have delayed convective initiation. Analysis of soundings indicates little difference in convective indices, suggesting that differences in convection may be attributed to choices in sub‐grid mixing parameters.

Turbulent mixing processes are important in determining the evolution of convective clouds,and the production of convective precipitation. However, the exact nature of these impacts remains uncertain due to limited observations. Model simulations show that assumptions made in parametrizing turbulence can have a marked effect on the characteristics of simulated clouds. This leads to significant uncertainty in forecasts from convection‐permitting numerical weather prediction (NWP) models. This contribution presents a comprehensive method to retrieve turbulence using Doppler weather radar to investigate turbulence in observed clouds. This method involves isolating the turbulent component of the Doppler velocity spectrum width, expressing turbulence intensity as an eddy dissipation rate, ϵ. By applying this method throughout large datasets of observations collected over the southern United Kingdom using the (0.28° beam‐width) Chilbolton Advanced Meteorological Radar (CAMRa), statistics of convective cloud turbulence are presented. Two contrasting case days are examined: a shallow “shower” case, and a “deep convection” case, exhibiting stronger and deeper updraughts. In our observations, ϵ generally ranges from 10−3 to 10−1 m2/s3, with the largest values found within, around and above convective updraughts. Vertical profiles of ϵ suggest that turbulence is much stronger in deep convection; 95th percentile values increase with height from 0.03 to 0.1 m2/s3, compared to approximately constant values of 0.02–0.03 m2/s3 throughout the depth of shower cloud. In updraught regions on both days, the 95th percentile of ϵ has significant (p < 10−3) positive correlations with the updraught velocity, and the horizontal shear in the updraught velocity, with weaker positive correlations with updraught dimensions. The ϵ‐retrieval method presented considers a very broad range of conditions, providing a reliable framework for turbulence retrieval using high‐resolution Doppler weather radar. In applying this method across many observations, the derived turbulence statistics will form the basis for evaluating the parametrization of turbulence in NWP models.

Convection-permitting ensembles have led to improved forecasts of many atmospheric phenomena. However, to fully utilize these forecasts the dependence of predictability on synoptic conditions needs to be understood. In this study, convective regimes are diagnosed based on a convective timescale which identifies the degree to which convection is in equilibrium with the large-scale forcing. Six convective cases are examined in a convection-permitting ensemble constructed using the Met Office Unified Model. The\ud ensemble members were generated using small-amplitude buoyancy perturbations added into the boundary layer, which can be considered to represent turbulent fluctuations close to the gridscale. Perturbation growth is shown to occur on different scales with an order of magnitude difference between the regimes (O(1 km) for cases closer to non-equilibrium convection and O(10 km) for cases closer to equilibrium convection). This difference reflects the fact that cell locations are essentially random in the equilibrium events after the first 12 h of the forecast, indicating a more rapid upscale perturbation growth compared to the non-equilibrium events. Furthermore, large temporal\ud variability is exhibited in all perturbation growth diagnostics for the nonequilibrium regime. Two boundary condition driven cases are also considered and show similar characteristics to the non-equilibrium cases, implying that caution is needed to interpret the timescale when initiation is not within the domain. Further understanding of perturbation growth within the different regimes could lead to a better understanding of where ensemble design improvements can be made beyond increasing the model resolution and could improve interpretation of forecasts.

Since 2013, the Met Office have run a 2.2 km horizontal gridlength version of the Unified Model (MetUM) as part of the National Oceanographic and Atmospheric Administration's Hazardous Weather Testbed Spring Forecasting Experiment. In this study, we perform high resolution MetUM simulations of the 20 May 2013 Oklahoma tornado outbreak at horizontal gridlengths between 2.2 km and 100 m. Here we present results showing that at 2.2 km gridlength the MetUM is able to simulate supercell-like storms whereas at O(100 m) gridlength it is able to simulate realistic-looking supercells with tornado-like vortices. This opens up the opportunity for using such simulations to highlight areas of enhanced tornado risk ahead of time.

Convection-permitting modelling has led to a step change in forecasting convective events. However, convection occurs within different regimes which exhibit different\ud forecast behaviour. A convective adjustment timescale can be used to distinguish between these regimes and examine their associated predictability. The convective adjustment timescale is calculated from radiosonde ascents and found to be consistent with that derived from convection-permitting model forecasts. The model-derived\ud convective adjustment timescale is then examined for three summers in the British Isles to determine characteristics of the convective regimes for this maritime region. Convection in the British Isles is predominantly in convective quasi-equilibrium with\ud 85%of convection having a timescale less than or equal to three hours. This percentage varies spatially with more non-equilibriumevents occurring in the south and southwest. The convective adjustment timescale exhibits a diurnal cycle over land. The nonequilibrium regime occurs more frequently at mid-range wind speeds and with winds\ud from southerly to westerly sectors. Most non-equilibrium convective events in the British Isles are initiated near large coastal orographic gradients or on the European\ud continent. Thus, the convective adjustment timescale is greatest when the location being examined is immediately downstream of large orographic gradients and decreases with distance from the convective initiation region. The dominance of convective quasiequilibrium conditions over the British Isles argues for the use of large-member\ud ensembles in probabilistic forecasts for this region.

Many factors, both mesoscale and larger scale, often come together in order for a particular convective initiation to take place. The authors describe a modeling study of a case from the Convective Storms Initiation Project (CSIP) in which a single thunderstorm formed behind a front in the southern United Kingdom. The key features of the case were a tongue of low-level high θw air associated with a forward-sloping split front (overrunning lower θw air above), a convergence line, and a “lid” of high static stability air, which the shower was initially constrained below but later broke through. In this paper, the authors analyze the initiation of the storm, which can be traced back to a region of high ground (Dartmoor) at around 0700 UTC, in more detail using model sensitivity studies with the Met Office Unified Model (MetUM). It is established that the convergence line was initially caused by roughness effects but had a significant thermal component later. Dartmoor had a key role in the development of the thunderstorm. A period of asymmetric flow over the high ground, with stronger low-level descent in the lee, led to a hole in a layer of low-level clouds downstream. The surface solar heating through this hole, in combination with the tongue of low-level high θw air associated with the front, caused the shower to initiate with sufficient lifting to enable it later to break through the lid.

With many operational centers moving toward order 1-km-gridlength models for routine weather forecasting, this paper presents a systematic investigation of the properties of high-resolution versions of the Met Office Unified Model for short-range forecasting of convective rainfall events. The authors describe a suite of configurations of the Met Office Unified Model running with grid lengths of 12, 4, and 1 km and analyze results from these models for a number of convective cases from the summers of 2003, 2004, and 2005. The analysis includes subjective evaluation of the rainfall fields and comparisons of rainfall amounts, initiation, cell statistics, and a scale-selective verification technique. It is shown that the 4- and 1-km-gridlength models often give more realistic-looking precipitation fields because convection is represented explicitly rather than parameterized. However, the 4-km model representation suffers from large convective cells and delayed initiation because the grid length is too long to correctly reproduce the convection explicitly. These problems are not as evident in the 1-km model, although it does suffer from too numerous small cells in some situations. Both the 4- and 1-km models suffer from poor representation at the start of the forecast in the period when the high-resolution detail is spinning up from the lower-resolution (12 km) starting data used. A scale-selective precipitation verification technique implies that for later times in the forecasts (after the spinup period) the 1-km model performs better than the 12- and 4-km models for lower rainfall thresholds. For higher thresholds the 4-km model scores almost as well as the 1-km model, and both do better than the 12-km model.

The development of NWP models with grid spacing down to 1 km should produce more realistic forecasts of convective storms. However, greater realism does not necessarily mean more accurate precipitation forecasts. The rapid growth of errors on small scales in conjunction with preexisting errors on larger scales may limit the usefulness of such models. The purpose of this paper is to examine whether improved model resolution alone is able to produce more skillful precipitation forecasts on useful scales, and how the skill varies with spatial scale. A verification method will be described in which skill is determined from a comparison of rainfall forecasts with radar using fractional coverage over different sized areas. The Met Office Unified Model was run with grid spacings of 12, 4, and 1 km for 10 days in which convection occurred during the summers of 2003 and 2004. All forecasts were run from 12-km initial states for a clean comparison. The results show that the 1-km model was the most skillful over all but the smallest scales (approximately