AbstractHigh‐resolution measurements from the Airborne Cloud‐Turbulence Observation System (ACTOS) during the Azores Stratocumulus Measurements of Radiation, Turbulence and Aerosols (ACORES) campaign are analysed for an investigation of the vertical profiles of microphysical properties and entrainment velocity (We) in marine stratocumulus clouds. The vertical profiles show the transition from the cloudy layer to free troposphere with nearly linear profiles of total water mixing ratio, liquid water potential temperature and virtual potential temperature, but the thickness of entrainment interfacial layer varies significantly. Sharp transitions of cloud microphysical and optical properties within a single horizontal flight leg are found in one stratocumulus cloud system. They seem to be related to the local environmental conditions, such as the wind shear and turbulent dissipation rate. We values estimated by three methods show consistent tendencies in general and are clearly related to the local environmental conditions, such as vertical shear of the horizontal wind and turbulence intensity. However, the magnitudes of We values differ by up to two orders of magnitude depending on the methods, which suggests that the estimation of We from in situ measurements is still a challenge. Analysis of the microphysical response to entrainment suggests that inhomogeneous mixing occurs dominantly. On the other hand, the analysis results for the clouds under more humid conditions indicate a higher likelihood of homogeneous mixing.

AbstractNumerical‐model‐based forecasts of precipitation exhibit poor skill over northern tropical Africa when compared with climatology‐based forecasts and with other tropical regions. However, as recently demonstrated, purely data‐driven forecasts based on spatio‐temporal dependences inferred from gridded satellite rainfall estimates show promise for the prediction of the 24‐hr precipitation occurrence rate in this region. The present work explores this potential further by advancing the statistical model and providing meteorological interpretations of the performance results. Advances include (a) the use of a recently developed correlation metric, the Coefficient of Predictive Ability (CPA), to identify predictors, (b) forecast evaluation with robust reliability diagrams and score decompositions, (c) a study domain over tropical Africa nested in a considerably enlarged spatio‐temporal domain to identify coherent propagating features, and (d) the introduction of a novel coherent‐linear‐propagation factor to quantify the coherence of propagating signals. The statistical forecast is compared with a climatology‐based benchmark, the European Centre for Medium‐Range Weather Forecasts (ECMWF) operational ensemble forecast, and a statistically postprocessed ensemble forecast. All methods show poor skill within the main rainbelt over northern tropical Africa, where differences in Brier scores between the different approaches are hardly statistically significant. However, the data‐driven forecast outperforms the other methods along the fringes of the rainbelt, where meridional rainfall gradients are large. The coherent‐linear‐propagation factor, in concert with metrics of convective available potential energy and convective instability, reveals that high stochasticity in the rainbelt limits predictability. At the fringes of the rainbelt, the data‐driven approach leverages coherent precipitation features associated with propagating tropical weather systems such as African Easterly Waves.

This study investigates the skill of the COSMO model (v5.7) at 1.1 km horizontal grid size in simulating the near‐surface foehn properties and evolution for five south foehn events and a 5‐year‐long climatology. A significant near‐surface cold bias is found during foehn, with an average bias of ‐3 K in the Rhine Valley in the five foehn cases, and of ‐1.8 K in the major northern foehn valleys in the 5‐year foehn climatology. The cold bias tends to be larger in the stronger and moister deep foehn events. Sensitivity experiments are carried out to examine the possible causes of the cold bias, including changes to the parameterization of the land‐atmosphere interaction, to the 1D turbulence parameterization, and to the horizontal grid spacing. Most sensitivity experiments have only a very minor impact on the cold bias, except for the model run with a horizontal grid spacing of 550 m. The 550 m COSMO run shows a reduced cold bias during foehn hours and also an improvement in the simulated foehn duration and northward foehn extent. By inspecting the vertical dimension, we found that the near‐surface cold bias downstream might partly originate upstream. A further contribution to the downstream cold bias is likely due to insufficient vertical mixing in the foehn flow. The latter is possibly enhanced in the 550 m model run, leading to a less stably stratified atmosphere in the lower few hundred meters of the atmosphere and a reduction of the reported model cold bias.This article is protected by copyright. All rights reserved.

AbstractThe operational Aeolus Level‐2B (L2B) Horizontal Line‐of‐Sight (HLOS) retrieved Rayleigh winds, produced by the European Space Agency (ESA), utilize ECMWF short‐term forecasts of temperature, pressure, and horizontal winds in the Rayleigh–Brillouin and M1 correction procedures. These model fields or backgrounds can contain ECMWF model‐specific errors, which may propagate to the retrieved Rayleigh winds. This study examines the sensitivity of the retrieved Rayleigh winds to the changes in the model backgrounds, and the potential benefit of using the same system, in this case the NOAA Finite‐Volume Cubed Sphere Global Forecast System (FV3GFS), for both the corrections and the data assimilation and forecast procedures.It is shown that the differences in the model backgrounds (FV3GFS‐ECMWF) can propagate through the L2B HLOS Rayleigh wind retrieval process, mainly the M1 correction, resulting in differences in the retrieved Rayleigh winds with mean and standard deviation of magnitude as large as 0.2 m s−1. The differences reach up to 0.4, 0.6, and 0.7 m/s for the 95th, 99th, and 99.5th percentiles of the sample distribution with maxima of ~1.4 m/s. The numbers of the large differences for the combined lower and upper 5th, 1st, and 0.5th percentile pairs are ~6100, 1220, and 610 between 2.5‐25 km height globally per day respectively. The ESA disseminated Rayleigh wind product (based on the ECMWF corrections) already shows a significant positive impact on the FV3GFS global forecasts (Garrett et al, 2022). In the observing system experiments (OSEs) performed, compared to the ESA Rayleigh winds, the use of the FV3GFS corrected Rayleigh winds lead to ~0.5% more Rayleigh winds assimilated in the lower troposphere and show enhanced positive impact on FV3GFS forecasts at the day 1‐10 range but limited to the Southern Hemisphere.This article is protected by copyright. All rights reserved.

Trapped lee waves, and resultant turbulent rotors downstream, present a hazard for aviation and land‐based transport. While high resolution numerical weather prediction models can represent such phenomena, there is currently no simple and reliable automated method for detecting the extent and characteristics of these waves in model output. Spectral transform methods have traditionally been used to detect and characterise regions of wave activity in model and observational data, however these methods can be slow and have their limitations. Machine learning (ML) techniques offer a new and potentially fruitful method of tackling this problem.This paper demonstrates that a deep learning model can be trained to accurately recognise and label coherent regions of lee waves from vertical velocity data on a single level from a high‐resolution NWP (numerical weather prediction) model. Using transfer learning, wave characteristics (wavelength, orientation and amplitude) can be extracted from the trained segmentation model. The use of synthetic wave fields with prescribed wave characteristics makes this transfer learning possible without the need to characterise real complex wave fields. Addition of noise to the synthetic data makes the models more robust when applied to more complex and noisy NWP data. The collection of trained models produced provides a valuable tool to investigate the prevalence and nature of lee wave activity, as well as a new way for forecasters to detect resolved waves. The deep learning model was more capable and quicker at detecting and characterising lee waves compared to a spectral technique.This work is just one example of how already established machine learning techniques can be used to detect and characterise complex weather phenomena from NWP model output and observational data, and how the careful use of synthetic data can reduce the requirements for large volumes of hand‐labelled training data for ML models.This article is protected by copyright. All rights reserved.

High‐resolution and accurate prediction of near‐surface weather parameters based on numerical weather prediction (NWP) models are essential for many downstream and real‐world applications. Traditional dynamical or statistical downscaling methods are insufficient to derive high‐resolution data from operational NWP forecasts, making it essential to devise new approaches. In recent years, an increasing number of researchers explored the implementations of deep learning (DL) based models for spatial downscaling, motivated by the similarity between the super‐resolution (SR) problem in computer vision (CV) and downscaling. Furthermore, while transformer‐based models have become state‐of‐the‐art models for many SR tasks, they are yet rarely applied for downscaling of weather forecasts or climate projections. This study adapted the transformer‐based models such as SwinIR and Uformer to downscale the temperature at 2 meters (T2m) and wind speed at 10 meters (WS10m) over Eastern Inner Mongolia, encompassing the area from 39.6°N to 46°N latitude and 111.6°E to 118°E longitude. We used the high‐resolution forecast (HRES) data from the European Centre for Medium‐range Weather Forecast (ECMWF) with a spatial resolution of 0.1° as the input and the gridded observation data, China Meteorological Administration (CMA) Land Data Assimilation System (CLDAS), at a spatial resolution of 0.01° as the target. Given that the models use observation data rather than a coarse‐grained version of forecast data as the target, the models accomplish both bias correction and spatial downscaling. The results demonstrate that the performance of the SwinIR and Uformer are superior to that of two convolutional neural networks (CNNs) based models (UNet and RCAN). Additionally, we introduced a novel module to extract features of varying resolutions from the high‐resolution topography data and applied a multi‐scale feature fusion module to merge features of different scales, contributing to further enhancement of the Uformer's performance.This article is protected by copyright. All rights reserved.

A theory for flow over gentle hills using a mixing length turbulence closure is developed to describe the transition from turbulent orographic form drag to gravity wave drag. It confirms that the first is associated with downstream sheltering, the second with upstream blocking and strong downslope winds. It shows that the altitude at which the incident flow needs to be taken to calculate the drag is the inner layer scale at which dissipation equilibrates disturbance advection. It also shows that the parameter that controls the transition, here a Richardson number, compares the mountain length to the altitude of the turning points above which the upward propagating gravity waves become evanescent. Our solutions are also used to show that the downslope winds penetrate well into the inner layer and that a good fraction of the drag is deposited in the inner layer: all of it in the neutral case, a large fraction in the intermediate cases when there are trapped lee waves, and even in stable situations without trapping part of the gravity wave drag is eroded in the inner layer. Some discussion on how to combine neutral and stratified effects in the parameterizationw of subgrid scale orography in large‐scale models is given.This article is protected by copyright. All rights reserved.

Abstract1976 was one of the most acute droughts in the UK, exceptional due to the compounding effects of low rainfall and hot summer temperatures. In this study, we provide a multi‐perspective view of the likelihood of a 1976‐like compound event occurring now and into the future. We find a high level of consistency in the messages emerging across a range of different approaches and climate modelling tools, from convection‐permitting climate projections to decadal hindcasts and global coupled model attribution ensembles. 1976 summer‐average temperatures remain, at the time of writing, amongst the highest on record, but with warming are becoming increasingly common. The 9‐month rainfall deficit to August 1976 was incredibly rare. Analysis here indicates that compound extremes like 1976 are expected to occur on timescales of hundreds to thousands of years in the present‐day climate, decreasing slightly into the future. The probability remains very small even if we account for favourable sea surface temperatures and atmospheric circulation that occurred in 1976. Similar but less severe events with a 1% chance of occurring in the present‐day are 5‐times more frequent from the 2040s under RCP8.5. The occurrence of such compound events is significantly (up to an order of magnitude) higher than expected if temperature and rainfall extremes occurred independently. In general, differences in likelihood estimates between approaches can begin to be understood from how dependence between variables is handled, differences in bias correction, and different levels of conditioning (i.e., the probability given particular atmospheric or ocean state). The appropriate choice of conditioning very much depends on the question being asked and its unconscious use may lead to apparent contradictions. Parallels can be drawn between 1976 and the recent summer 2022, and results here suggest that with hotter summers we should be prepared for more severe droughts like 1976 in the future.This article is protected by copyright. All rights reserved.

The Land surface Interactions with the Atmosphere over the Iberian Semi‐arid Environment (LIAISE) campaign was conducted in July 2021 primarily to investigate the role of irrigation in modulating the boundary layer evolution in the Catalan region of northeastern Spain. Contrasts in near‐surface meteorological parameters and boundary layer thermodynamic profiles at an irrigated and rainfed (arid) site were established during the morning transition. Evapotranspriation dominated the flux partitioning at the irrigated site (Bowen ratio of 0.07–1.1), whilst sensible heat flux dominated at the rainfed (arid) site (Bowen ratio greater than 10.0). The cumulative evapotranspiration during July 2021 was a factor of 10 greater at the irrigated site than at the rainfed (arid) site. The presence of irrigation was shown to modulate the vertical gradients of turbulence, temperature, and moisture. Irrigation is shown to have a significant effect on the development of the boundary layer including during the morning transition. The morning transition mean buoyancy flux was 2.8 times smaller at the irrigated site (1.1 m2 s−2) compared with the rainfed (arid) site (3.1 m2 s−2), with a resultant delay in the near‐surface buoyancy‐flux crossover time (30‐min to 90‐min) at the irrigated site. At the start of the morning transition (sunrise), the average screen‐level (50 m) temperature was ‐1.2 K (‐1.9 K) colder at the irrigated site relative to the rainfed (arid) site. The colder temperatures at sunrise at the irrigated site are predominately the result of colder boundary layer thermodynamic profile from the previous day. At the end of the morning transition (convective onset), temperature differences between the two sites extend through much of the boundary layer and increased in magnitude. The average screen‐level (50 m) temperature difference was ‐3.6 K (‐2.4 K) colder at the irrigated site relative to the rainfed (arid) site. There was considerable day‐to‐day variability in temperature contrasts at a regional level (‐2.4 to ‐6.0 K).This article is protected by copyright. All rights reserved.

AbstractThe tropical Pacific plays an important role in modulating the global climate through its prevailing sea‐surface temperature spatial structure and dominant climate modes like El Niño–Southern Oscillation (ENSO), the Madden–Julian Oscillation (MJO), and their teleconnections. These modes of variability, including their oceanic anomalies, are considered to provide sources of prediction skill on subseasonal timescales in the Tropics. Therefore, this study aims to examine how assimilating in situ ocean observations influences the initial ocean sea‐surface temperature (SST) and mixed‐layer depth (MLD) and their subseasonal forecasts. We analyze two subseasonal forecast systems generated with the European Centre for Medium‐Range Weather Forecasts (ECMWF) Integrated Forecast System (IFS), where the ocean states were initialized using two Observing‐System Experiment (OSE) reanalyses. We find that the SST differences between forecasts with and without ocean data assimilation grow with time, resulting in a reduced cold‐tongue bias when assimilating ocean observations. Two mechanisms related to air–sea coupling are considered to contribute to this growth of SST differences. One is a positive feedback between the zonal SST gradient, pressure gradient, and surface wind. The other is the difference in Ekman suction and mixing at the Equator due to surface wind‐speed differences. While the initial mixed‐layer depth (MLD) can be improved through ocean data assimilation, this improvement is not maintained in the forecasts. Instead, the MLD in both experiments shoals rapidly at the beginning of the forecast. These results emphasize how initialization and model biases influence air–sea interaction and the accuracy of subseasonal forecasts in the tropical Pacific.