Fine-Scale Structures in the Mid-Level Eyewall of Super Typhoon Rammasun (2014) Simulated With the WRF-LES Framework

It has been numerically demonstrated that the turbulence above the boundary is important to tropical cyclone intensification and rapid intensification, but the three-dimensional structures of the sub-grid-scale (SGS) eddy have not been revealed due to the lack of observational data. In this study, two numerical simulations of Super Typhoon Rammasun (2014) were conducted with the Advanced Weather Research and Forecast (WRF) model by incorporating the large-eddy simulation (LES) technique, in which the enhanced eyewall convection and the process of rapid intensification are captured. Consistent with previous observational studies, the strong turbulent kinetic energy (TKE) is found throughout the whole eyewall inside of the radius of maximum wind in both experiments. The simulations indicate that the strong TKE is associated with horizontal rolls with the horizontal extent of 2–4 km, which are aligned azimuthally in the intense eyewall convection. It is indicated that the three-dimensional structures of the SGS eddy can be simulated with the vertical grid spacing of ∼100 m when the horizontal grid spacing is 74 m. It is suggested that there is considerable turbulence associated with azimuthally-aligned horizontal rolls in the mid-level eyewall of tropical cyclone.

Evaluation of Mei-yu heavy-rainfall quantitative precipitation forecasts in Taiwan by a cloud-resolving model for three seasons of 2012–2014

In this study, the performance of quantitative precipitation forecasts (QPFs) by the Cloud-Resolving Storm Simulator (CReSS) in Taiwan, at a horizontal grid spacing of 2.5 km and a domain size of 1500×1200 km2, in the range of 1–3 d during three Mei-yu seasons (May–June) of 2012–2014 is evaluated using categorical statistics, with an emphasis on heavy-rainfall events (≥100 mm per 24 h). The categorical statistics are chosen because the main hazards are landslides and floods in Taiwan, so predicting heavy rainfall at the correct location is important. The overall threat scores (TSs) of QPFs for all events on day 1 (0–24 h) are 0.18, 0.15, and 0.09 at thresholds of 100, 250, and 500 mm, respectively, and indicate considerable improvements at increased resolution compared to past results and 5 km models (TS < 0.1 at 100 mm and TS ≤ 0.02 at 250 mm). Moreover, the TSs are shown to be higher and the model more skillful in predicting larger events, in agreement with earlier findings for typhoons. After classification based on observed rainfall, the TSs of day − 1 QPFs for the largest 4 % of events by CReSS at 100, 250, and 500 mm (per 24 h) are 0.34, 0.24, and 0.16, respectively, and can reach 0.15 at 250 mm on day 2 (24–48 h) and 130 mm on day 3 (48–72 h). The larger events also exhibit higher probability of detection and lower false alarm ratio than smaller ones almost without exception across all thresholds. With the convection and terrain better resolved, the strength of the model is found to lie mainly in the topographic rainfall in Taiwan rather than migratory events that are more difficult to predict. Our results highlight the crucial importance of cloud-resolving capability and the size of fine mesh for heavy-rainfall QPFs in Taiwan.

Exploring the Usefulness of Downscaling Free Forecasts from the Warn-on-Forecast System

The National Severe Storms Lab (NSSL) Warn-on-Forecast System (WoFS) is an experimental real-time rapidly-updating convection-allowing ensemble that provides probabilistic short-term thunderstorm forecasts. This study evaluates the impacts of reducing the forecast model horizontal grid spacing Δx from 3 km to 1.5 km on the WoFS deterministic and probabilistic forecast skill, using eleven case days selected from the 2020 NOAA Hazardous Weather Testbed (HWT) Spring Forecasting Experiment (SFE). Verification methods include (i) subjective forecaster impressions; (ii) a deterministic object-based technique that identifies forecast reflectivity and rotation track storm objects as contiguous local maxima in the composite reflectivity and updraft helicity fields, respectively, and matches them to observed storm objects; and (iii) a recently developed algorithm that matches observed mesocyclones to mesocyclone probability swath objects constructed from the full ensemble of rotation track objects. Reducing Δx fails to systematically improve deterministic skill in forecasting reflectivity object occurrence, as measured by critical success index (CSIDET), a metric that incorporates both probability of detection (PODDET) and false alarm ratio (FARDET). However, compared to the Δx = 3 km configuration, the Δx = 1.5 km WoFS shows improved mid-level mesocyclone detection, as evidenced by its statistically significant (i) higher CSIDET for deterministic mid-level rotation track objects and (ii) higher normalized area under the performance diagram curve (NAUPDC) score for probability swath objects. Comparison between Δx = 3 km and Δx = 1.5 km reflectivity object properties reveals that the latter have 30% stronger mean updraft speeds, 17% stronger median 80-m winds, 67% larger median hail diameter, and 28% higher median near-storm-maximum 0-3 km storm-relative helicity.

Horizontal Grid Spacing Comparison Among Random Forest Algorithms to Nowcast Cloud-to-Ground Lightning Occurrence

This paper discusses the use of Random Forest (RF), a popular Machine Learning (ML) algorithm, to perform spatially explicit nowcasting of cloud-to-ground lightning occurrence. An application to the Italian territory and the surrounding seas is presented. Specifically, 1-hour ahead lightning occurrences over the months of August, September and October from 2017 to 2019 have been modelled using a dataset including geo-environmental features. Results obtained with three different spatial resolutions have been compared, for nowcasting both positive and negative strokes. The features’ importance resulting from the best RF models showed how datadriven models are able to identify the relationships between meteorological variables, in agreement with previous physically based knowledge of the phenomenon. The encouraging results obtained in terms of forecasting accuracy support the idea to use ML-based algorithms in early warning procedures for disaster risk management.

On the application and grid-size sensitivity of the urban dispersion model CAIRDIO for the simulation of a real city and realistic meteorology

There is a gap between the need for city-wide air-quality simulations considering the intra-urban variability and mircoscale dispersion features and the computational capacities that conventional urban microscale models require. This gap can be bridged by targeting model applications on the gray zone situated between the mesoscale and large-eddy scale. The urban dispersion model CAIRDIO is a new contribution to the class of computational-fluid dynamics models operating in this scale range. It uses a diffuse-obstacle boundary method to represent buildings as physical obstacles at gray-zone resolutions in the order of tens of meters. The main objective of this approach is to find an acceptable compromise between computationally inexpensive grid sizes for spatially comprehensive applications and the required accuracy in the description of building and boundary-layer effects. For this purpose, CAIRDIO is applied in dispersion simulation of black carbon and particulate matter for an entire mid-size city using an uniform horizontal resolution of 40 m in this paper. For evaluation, the simulation results are compared with measurements from 5 operational air monitoring stations, which are representative for the urban background and high-traffic roads, respectively. Moreover, the comparison includes the mesoscale host simulation, which provides the boundary conditions. The temporal variability of the concentration measurements at the background sites was largely influenced only by the characteristics of the mixing layer. As a consequence, the model results were not significantly dependent on spatial resolution, so that the mesoscale simulation also performed reasonably well. At the traffic sites, however, concentrations were in addition markedly influenced by the proximity to road-traffic sources and the surrounding building environment. Here, the mesoscale simulation indiscriminately reproduced almost the same urban-background profiles, which resulted in a large positive model bias. On the other hand, the CAIRDIO simulation was able to respond to the significantly amplified diurnal variability with its pronounced rush-hour peaks. This resulted in a consistent improvement of the model deviation to mea- surements compared to the mesoscale simulation. Nevertheless, discrepancies to measurements remain in the 40 m-CAIRDIO simulation, e.g., an underestimation of peak concentrations at two traffic sites inside narrow street canyons. To further research resolution sensitivity, the horizontal grid spacing of locally nested CAIRDIO domains is refined down to 5 m. While for the street canyons the representation of peak concentrations can be improved using horizontal grid spacings of up to 10 m, no further improvements beyond this resolution can be observed. This suggests that the too low peak concentrations with the default grid spacing of 40 m result from an inadequate representation of the traffic emissions inside narrow street canyons. If the total gain in accuracy due to the grid refinements is put in relation to the remaining model error, the improvements are only modest. In conclusion, the proposed gray-scale modeling is a promising downscaling approach for urban air-quality applications. Nevertheless, the results also show that aspects other than the actual resolution of flow patterns and numerical effects can determine the simulations at the urban microscale.

Evaluation of a forest parameterization to improve boundary layer flow simulations over complex terrain

We evaluate the influence of a forest parametrization on the simulation of the boundary layer flow over moderate complex terrain in the context of the Perdigão 2017 field campaign. The numerical simulations are performed using the Weather research and forecasting model using its large eddy simulation mode (WRF-LES). The short-term high resolution (40 m horizontal grid spacing) and long-term (200 m horizontal grid spacing) WRF-LES are evaluated for an integration time of 12 hours and 1.5 months, respectively, with and without forest parameterization. The short-term simulations focus on low-level jet events over the valley, while the long-term simulations cover the whole intensive observation period (IOP) of the field campaign. The results are validated using lidar and meteorological tower observations. The mean diurnal cycle during the IOP shows a significant improvement of the along-valley wind speed and the wind direction when using the forest parametrization. However, the drag imposed by the parametrization results in an underestimation of the cross-valley wind speed, which can be attributed to a poor representation of the land surface characteristics. The evaluation of the high-resolution WRF-LES shows a positive influence of the forest parametrization on the simulated winds in the first 500 m above the surface.

High-Definition Hurricanes: Improving Forecasts with Storm-Following Nests

To forecast tropical cyclone (TC) intensity and structure changes with fidelity, numerical weather prediction models must be “high definition”, i.e., horizontal grid spacing ≤ 3 km, so that they permit clouds and convection and resolve sharp gradients of momentum and moisture in the eyewall and rainbands. However, resolutions in operational global models remain too coarse to accurately predict these structures that are critical to TC intensity. Storm-following nests are a solution to this problem because they are computationally efficient at fine resolutions, providing a practical approach to improve TC intensity forecasts. Under the Hurricane Forecast Improvement Program, the operational Hurricane Weather Research and Forecasting (HWRF) system was developed to include telescopic, storm-following nests for a single TC per model integration. Subsequently, HWRF evolved into a state-of-the-art tool for TC predictions around the globe, although its single-storm nesting approach does not adequately simulate TC-TC interactions as they are observed. Basin-scale HWRF (HWRF-B) was developed later with a multi-storm nesting approach to improve the simulation of TC-TC interactions by producing high-resolution forecasts for multiple TCs simultaneously. In this study, the multi-storm nesting approach in HWRF-B was compared with a single-storm nesting approach using an otherwise identical model configuration. The multi-storm approach demonstrated TC intensity forecast improvements, including more realistic TC-TC interactions. Storm-following nests developed in HWRF and HWRF-B will be foundational to NOAA’s next-generation hurricane application in the Unified Forecast System.

Microphysical processes producing high ice water contents (HIWCs) in tropical convective clouds during the HAIC-HIWC field campaign: dominant role of secondary ice production

High ice water content (HIWC) regions in tropical deep convective clouds, composed of high concentrations of small ice crystals, were not reproduced by Weather Research and Forecasting (WRF) model simulations at 1-km horizontal grid spacing using four different bulk microphysics schemes (i.e., the WRF single‐moment 6‐class microphysics scheme (WSM6), the Morrison scheme and the Predicted Particle Properties (P3) scheme with one- and two-ice options) for conditions encountered during the High Altitude Ice Crystals (HAIC)-HIWC experiment. Instead, overestimates of radar reflectivity and underestimates of ice number concentrations were realized. To explore formation mechanisms for large numbers of small ice crystals in tropical convection, a series of quasi-idealized WRF simulations varying the model resolution, aerosol profile, and representation of secondary ice production (SIP) processes are conducted based on an observed radiosonde released at Cayenne during the HAIC-HIWC field campaign. The P3 two-ice scheme, which has two “free” ice categories to represent all ice-phase hydrometeors, is used. Regardless of the horizontal grid spacing or aerosol profile used, without including SIP processes the model produces total ice number concentrations about two orders of magnitude less than observed at −10 °C and about an order of magnitude less than observed at −30 °C, but slightly overestimates the total ice number concentrations at −45 °C. Three simulations including one of three SIP mechanisms separately (i.e., the Hallett-Mossop mechanism, fragmentation during ice–ice collisions, and shattering of freezing droplets) also do not replicate observed HIWCs, with the results of the simulation including shattering of freezing droplets most closely resembling the observations. The simulation including all three SIP processes successfully produces HIWC regions at all temperature levels remarkably consistent with the observations in terms of ice number concentrations and radar reflectivity, which is not replicated using the original P3 two-ice scheme. This simulation shows that primary ice production plays a key role in generating HIWC regions at t < 30 min at temperatures < −40 °C, shattering of freezing droplets dominates ice particle production in HIWC regions at temperatures > − 15 °C during the early stage of convection, and fragmentation during ice–ice collisions dominates at temperatures > −15 °C during the later stage of convection and at temperatures < −20 °C over the whole convection period. This study confirms the dominant role of SIP processes in the formation of numerous small crystals in HIWC regions.

Model inter-comparison for short-range forecasts over the southern African domain

Numerical weather prediction (NWP) models have been increasing in skill and their capability to simulate weather systems and provide valuable information at convective scales has improved in recent years. Much effort has been put into developing NWP models across the globe. Representation of physical processes is one of the critical issues in NWP, and it differs from one model to another. We investigated the performance of three regional NWP models used by the South African Weather Service over southern Africa, to identify the model that produces the best deterministic forecasts for the study domain. The three models – Unified Model (UM), Consortium for Small-scale Modelling (COSMO) and Weather Research and Forecasting (WRF) – were run at a horizontal grid spacing of about 4.4 km. Model forecasts for precipitation, 2-m temperature, and wind speed were verified against different observations. Snow was evaluated against reported snow records. Both the temporal and spatial verification of the model forecasts showed that the three models are comparable, with slight variations. Temperature and wind speed forecasts were similar for the three different models. Accumulated precipitation was mostly similar, except where WRF captured small rainfall amounts from a coastal low, while it over-estimated rainfall over the ocean. The UM showed a bubble-like shape towards the tropics, while COSMO cut-off part of the rainfall band that extended from the tropics to the sub-tropics. The COSMO and WRF models simulated a larger spatial coverage of precipitation than UM and snow-report records.

Comparing partial and continuously cycling ensemble Kalman filter data assimilation systems for convection-allowing ensemble forecast initialization

Several limited-area 80-member ensemble Kalman filter (EnKF) data assimilation systems with 15-km horizontal grid spacing were run over a computational domain spanning the conterminous United States (CONUS) for a 4-week period. One EnKF employed continuous cycling, where the prior ensemble was always the 1-h forecast initialized from the previous cycle’s analysis. In contrast, the other EnKFs used a partial cycling procedure, where limited-area states were discarded after 12 or 18 h of self-contained hourly cycles and re-initialized the next day from global model fields. “Blended” states were also constructed by combining large scales from global ensemble initial conditions (ICs) with small scales from limited-area continuously cycling EnKF analyses using a low-pass filter. Both the blended states and EnKF analysis ensembles initialized 36-h, 10-member ensemble forecasts with 3-km horizontal grid spacing. Continuously cycling EnKF analyses initialized ~1–18-h precipitation forecasts that were comparable to or somewhat better than those with partial cycling EnKF ICs. Conversely, ~18–36-h forecasts with partial cycling EnKF ICs were comparable to or better than those with unblended continuously cycling EnKF ICs. However, blended ICs yielded ~18–36-h forecasts that were statistically indistinguishable from those with partial cycling ICs. ICs that more closely resembled global analysis spectral characteristics at wavelengths > 200 km, like partial cycling and blended ICs, were associated with relatively good ~18–36-h forecasts. Ultimately, findings suggest that EnKFs employing a combination of continuous cycling and blending can potentially replace the partial cycling assimilation systems that currently initialize operational limited-area models over the CONUS without sacrificing forecast quality.