Search

This study proposes a weighted composite approach for grey relational analysis (GRA) that utilizes a numerical weather prediction (NWP) and support vector machine (SVM). The approach is optimized using an improved grey wolf optimization (IGWO) algorithm. Initially, the dimension of NWP data is decreased by t-distributed stochastic neighbor embedding (t-SNE), then the weight of sample coefficients is calculated by entropy-weight method (EWM), and the weighted grey relational of data points is calculated for different weather numerical time series data. At the same time, a new weighted composite grey relational degree is formed by combining the weighted cosine similarity of NWP values of the historical day and to be measured day. The SVM's regression power prediction model is constructed by the time series data. To improve the accuracy of the system's predictions, the grey relational time series data is chosen as the input variable for the SVM, and the influence parameters of the ideal SVM are discovered using the IGWO technique. According to the simulated prediction and analysis based on NWP, it can be observed that the proposed method in this study significantly improves the prediction accuracy of the data. Specifically, evaluation metrics such as root mean squared error (RMSE), regression correlation coefficient (r2), mean absolute error (MAE) and mean absolute percent error (MAPE) all show corresponding enhancements, while the computational burden remains relatively low. [ABSTRACT FROM AUTHOR]

This two‐part study presents a comprehensive analysis of (1) the vortex structure, (2) the inner core planetary boundary layer (PBL) wind profile, and (3) the overland surface winds of Typhoon Faxai(2019) during landfall observationally (Part I) and in high‐resolution Weather Research and Forecasting Model (WRF) simulations (Part II). Part II presents a framework for evaluating the wind field in the WRF simulations of Typhoon Faxai (2019) by comparing them with the observations presented in Part I. First, WRF simulations with two different surface roughness z0$$ {z}_0 $$ tables are presented. The simulation with the default z0$$ {z}_0 $$ table in WRF is shown to largely overestimate the surface wind speed. The overestimation is largely mitigated by using larger z0$$ {z}_0 $$ based on the climatology for urban and forest areas in Japan. Next, using the modified z0$$ {z}_0 $$ table, WRF simulations with three different PBL parametrizations, namely, Mellor–Yamada–Janjić, Yonsei University, and Mellor–Yamada–Nakanishi–Niino schemes are presented, and their impacts on the simulated vortex structure, inner core PBL wind profile, and overland surface winds are evaluated. In particular, the simulated surface winds are in good agreement with the observations outside the inner core of Typhoon Faxai. However, during the passage of the eyewall, the maximum surface winds are underestimated in the simulations. We show that the underestimation of the inertial stability, associated with the excessively large inner core of Typhoon Faxai in the simulations, led to the overestimation of the eyewall PBL jet height. This, in turn, caused the underestimation of the peak surface winds, as based on the empirical model of the wind reduction factor F2 km$$ {F}_{2\;\mathrm{km}} $$ proposed in Part I. [ABSTRACT FROM AUTHOR]

Accurate short-term solar irradiance forecasting is crucial for the efficient operation of solar energy-driven photovoltaic (PV) power plants. In this research, we introduce a novel hybrid ensemble forecasting model that amalgamates the strengths of machine learning tree-based models and deep learning neuron-based models. The hybrid ensemble model integrates the interpretability of tree-based models with the capacity of neuron-based models to capture complex temporal dependencies within solar irradiance data. Furthermore, stacking and voting ensemble strategies are employed to harness the collective strengths of these models, significantly enhancing the prediction accuracy. This integrated methodology is enhanced by incorporating pixels from satellite images provided by the European Organization for the Exploitation of Meteorological Satellites (EUMETSAT). These pixels are converted into structured data arrays and employed as exogenous inputs in the algorithm. The primary objective of this study is to improve the accuracy of short-term solar irradiance predictions, spanning a forecast horizon up to 6 h ahead. The incorporation of EUMETSAT satellite image pixel data enables the model to extract valuable spatial and temporal information, thus enhancing the overall forecasting precision. This research also includes a detailed analysis of the derivation of the GHI using satellite images. The study was carried out and the models tested across three distinct locations in Austria. A detailed comparative analysis was carried out for traditional satellite (SAT) and numerical weather prediction (NWP) models with hybrid models. Our findings demonstrate a higher skill score for all of the approaches compared to a smart persistent model and consistently highlight the superiority of the hybrid ensemble model for a short-term prediction window of 1 to 6 h. This research underscores the potential for enhanced accuracy of the hybrid approach to advance short-term solar irradiance forecasting, emphasizing its effectiveness at understanding the intricate interplay of the meteorological variables affecting solar energy generation worldwide. The results of this investigation carry noteworthy implications for advancing solar energy systems, thereby supporting the sustainable integration of renewable energy sources into the electrical grid. [ABSTRACT FROM AUTHOR]

This two‐part study presents a detailed analysis of (a) the vortex structure, (b) the inner core planetary boundary layer (PBL) wind profile, and (c) the overland surface winds of Typhoon Faxai (2019) during landfall observationally (Part I) and in high‐resolution Weather Research and Forecasting model simulations (Part II). Part I presents a comprehensive observational analysis of (a), (b), and (c) of Typhoon Faxai (2019) based on radar dual‐Doppler analysis and spatially dense surface wind observations over the Kanto region of Japan. An empirical model for the wind reduction factor F2 km$$ {F}_{2\;\mathrm{km}} $$ is proposed and validated with surface observations and radar velocity–azimuth display analyses at Narita and Haneda airports, both of which were affected by the passage of Faxai's inner core. Two distinct PBL jets associated with the inner rain band and the eyewall were observed over Narita Airport, which was located to the right of Faxai's track. Based on the empirical model of F2 km$$ {F}_{2\;\mathrm{km}} $$, it is found that the maximum surface wind speed observed at Narita Airport was caused by the peak of F2 km$$ {F}_{2\;\mathrm{km}} $$ due to a strong PBL jet associated with the eyewall located as close as ∼$$ \sim $$0.3 km from the ground. In addition, the wind swath analysis shows that a large area far inland on the right side of Faxai's track experienced strong surface 10‐min winds due to the combination of the passage of the eyewall PBL jet and the small surface roughness length around agricultural land such as paddy fields and farm fields. [ABSTRACT FROM AUTHOR]

Two methods are widely used to assess the impact of observations in global numerical weather prediction (NWP): data denial experiments (DDEs) and the forecast sensitivity‐based observation impact (FSOI) method. A DDE measures the impact on forecast accuracy of removing an observation type from the system, whereas FSOI estimates the amount by which an observation type reduces the short‐range forecast error within a system containing all observation types. This article describes the second part of a two‐part study. In the first part, the theory behind DDE and FSOI metrics was presented and then applied to a simple model with two state variables, all in the context of optimal data assimilation (DA), for which the error covariances used in the DA system match reality. The article showed why and under what conditions the DDE and FSOI metrics give different results, even for an optimal DA system. In this second part, we extend the theory to suboptimal systems, and specifically to systems that are suboptimal in their specification of observation errors, and then apply it to a very simple model, in this case with one state variable. As expected, DDE impacts are reduced when the system is suboptimal. By contrast, relative FSOI impacts (i.e. relative to those of other observation types) increase for an observation type for which the errors are underestimated. This gives the erroneous impression that the change in assumed errors has led to an improvement, whereas the opposite is the case. These results provide some insights into the interpretation of FSOI results from a suboptimal DA system. [ABSTRACT FROM AUTHOR]

In this article, a machine‐learning‐based tool for calibrating numerical forecasts of near‐surface air temperature is proposed. The study area covers Poland representing a temperate type of climate with transitional features and highly variable weather. The direct output of numerical weather prediction (NWP) models is often biased and needs to be adjusted to observed values. Forecasters have to reconcile forecasts from several NWP models during their operational work. As the proposed method is based on deterministic forecasts from three short‐range limited‐area models (ALARO, AROME and COSMO), it can support them in their decision‐making process. Predictors include forecasts of weather elements produced by the NWP models at synoptic weather stations across Poland and station‐embedded data on ambient orography. The Random Forests algorithm (RF) has been used to produce bias‐corrected forecasts on a test set spanning one year. Its performance was evaluated against the NWP models, a linear combination of all predictors (multiple linear regression, MLR) as well as a basic Artificial Neural Network (ANN). Detailed evaluation was done to identify potential strengths and weaknesses of the model at the temporal and spatial scale. The value of RMSE of a forecast obtained by the RF model was 11% and 27% lower compared to the MLR model and the best‐performing NWP model respectively. The ANN model turned out to be even superior, outperforming RF by around 2.5%. The greatest improvement occurred for warm bias during the nighttime from July to September. The largest difference in forecast accuracy between RF and ANN appeared for temperature drops inApril nights. Poor performance of RF for extreme temperature ranges may be suppressed by training the model on forecast error instead of observed values of the variable. [ABSTRACT FROM AUTHOR]

The European Space Agency's Aeolus mission, launched in August 2018, provides the first global horizontal line‐of‐sight wind profile measurements. Previous studies have shown that Aeolus winds in global forecast systems improve the overall forecast skill, especially in the upper tropospheric tropics and in other data‐sparse regions. In this study, we use a series of observing system experiments with the latest version of the reprocessed Aeolus wind product (2B11) to better characterize the locations and drivers of improved skill from Aeolus with Environment and Climate Change Canada's Global Deterministic Prediction System. Observing system experiments that test the impact of Aeolus winds and the impact of all operational wind observations are carried out, covering the period summer 2019 and winter 2019–2020. Assimilation of operational winds improves the tropospheric wind forecast over the Tropics by a reduction of 8% in the forecast error, and adding the Aeolus winds to the assimilations results in an extra 0.7–0.9%. Aeolus wind impacts are improvements are 0.7–0.9% for the Arctic, and 0.4–0.6% over the Northern and Southern Hemisphere extratropics. The scale dependence of these impacts is investigated using spatial spectra (spherical harmonic decomposition). The improvement is quantified using the difference of the 250 hPa kinetic energy forecast error spectra between experiments. The operational winds largely improve the forecast of planetary scale to intermediate scale for spherical wave numbers between 1 and 20 in the short‐range forecasts. The operational wind impact decreases as the forecast lead time increases. On the other hand, the impact of Aeolus is mostly seen in the intermediate to large scale range with a peak around spherical wave number 8. The Aeolus‐related improvement around this wave number increases with forecast lead time. This analysis suggests that Aeolus winds provide estimates of the wind state that are valuable and complementary to that provided from current operational winds. [ABSTRACT FROM AUTHOR]

In space geodetic data analysis, improving tropospheric delay modelling is motivated by the correlation between tropospheric zenith total delay (ZTD) and the station height. The gradients are correlated with the horizontal displacements. In the microwave techniques such as GPS, DORIS, and VLBI, the tropospheric delay effects are correlated over the collocation sites. This correlation allows for the estimation of common tropospheric parameters, often referred to as tropospheric ties. These tropospheric ties provide valuable complementary information for the computation of the terrestrial reference frame (TRF) and have been the subject of investigation in many studies. In this study, we investigate the effects of tropospheric ties on the daily TRF combination at the observation level using a batch least-squares estimation. The observations of GPS, DORIS, and VLBI were collected from 06 May 2014 to 20 May 2014 during the CONT14 campaign of VLBI. The tropospheric delay and gradient ties are computed using different numeric weather prediction (NWP) data sets provided by ECMWF and NCEP. We examined different levels of tropospheric ties 0.01, 5, and 10 mm for ZTD and 0.001, 0.5, and 1.0 mm for gradients in the combination of techniques. The results show that the combined solution with tropospheric ties derived from the four NWP data sets does not exhibit significant differences. For VLBI, the repeatability of station coordinates and network scale were found to be improved by around 20% and 30%, respectively. The stronger tropospheric ties show a higher improvement in VLBI baseline repeatability. However, applying tropospheric ties at GPS-DORIS collocation sites does not significantly affect the repeatability of station coordinates and network scale. Both ZTD and gradient ties enhance the repeatability of polar motion components in EOPs, while no observable contribution is observed for dUT1 and celestial pole offsets. [ABSTRACT FROM AUTHOR]

Global wind profiles provided by the satellite mission Aeolus are an important recent supplement to the Global Observing System. This study investigates the impact of Aeolus horizontal line‐of‐sight wind observations in the operational global assimilation and forecasting system of Deutscher Wetterdienst that is based on the Icosahedral Nonhydrostatic (ICON) model. For this purpose, an observing system experiment was conducted and evaluated for a 3‐month period from July 2020 to October 2020. The Aeolus Rayleigh clear and Mie cloudy data quality and consistency were derived from observation minus background statistics. To correct for an altitude‐dependent bias, a model‐based bias correction scheme has been implemented. Comparisons of the systematic changes in the analysis and the respective forecasts provide an overview of the overall impact of the Aeolus horizontal line‐of‐sight wind assimilation in ICON. Increased influence of Aeolus wind profiles is found in jet regimes (e.g., amplification of the zonal wind component), around large‐scale circulation systems, and convectively active areas in the Tropics. The reduction in forecast error is largest in the tropical upper troposphere and stratosphere, as well as in the mid and upper troposphere of the Southern Hemisphere. The Northern Hemisphere shows a somewhat smaller but still beneficial impact of Aeolus observations. The verification with other conventional observations shows a mean relative reduction in short‐range forecast error between 0.1% and 0.6% in the Northern Hemisphere and up to 1.6% in the Tropics and the Southern Hemisphere. When verifying against the European Centre for Medium‐Range Weather Forecast Reanalysis v5, forecast errors of zonal wind, temperature, and geopotential up to 5 days lead time are reduced by 2–4% on global average and up to 5–8% around the tropical tropopause. [ABSTRACT FROM AUTHOR]

This study investigates the impact of assimilating Aircraft Meteorological Data Relay (AMDAR) observations on the prediction of two typhoons, Nesat and Haitang (2017), using the Gridpoint Statistical Interpolation (GSI) assimilation system and the Weather Research and Forecasting (WRF) model. Two quality control (QC) methods, Minimum Covariance Determinant (MCD) and one-class Support Vector Machine (OCSVM), were employed to perform QC on the AMDAR observations before data assimilation. The QC results indicated that both methods significantly reduced kurtosis, skewness, and discrepancies between the AMDAR data and the reanalysis data. The data distribution after applying the MCD-QC method exhibited a closer resemblance to a Gaussian distribution. Four numerical experiments were conducted to assess the impact of different AMDAR data qualities on typhoon forecasting, including a control experiment without data assimilation (EXP-CNTL), assimilating all AMDAR observations (EXP-RAW), assimilating observations after applying MCD-QC (EXP-MCD), and assimilating observations after applying OCSVM-QC (EXP-SVM). The results demonstrated that using AMDAR data in assimilation improved the track and intensity prediction of the typhoons. Furthermore, utilizing QC before assimilation enhanced the performance of track forecasting prediction, with EXP-MCD showing the best performance. As for intensity prediction, the three assimilation experiments exhibited varying strengths and weaknesses at different times, with EXP-MCD showing smaller intensity forecast errors on average. [ABSTRACT FROM AUTHOR]

With the advent of the low-carbon era, wind power has become an indispensable energy source. Accurate day-ahead wind speed forecast is crucial for the power system to absorb wind power. Due to the influence of the spatiotemporal resolution and the error of forecasting itself, there is a certain error between the original wind speed of numerical weather prediction and the actual wind speed. Aiming to minimize this error as much as possible, this paper advocates using multivariable space-time information to jointly correct the wind speed in numerical weather prediction. Firstly, the correlation analysis experiments are carried out to demonstrate the feasibility of the idea. Then, the multivariable space-time experiment based on convolutional long short-term memory network is carried out, which greatly reduced the initial wind speed error in numerical weather prediction. At the same time, various methods are used for comparison. The experimental results show that the proposed method reduces the mean absolute error of the numerical weather prediction by 41.13% ~ 77.70% and reduces the root mean square error of the numerical weather prediction by 37.30% ~ 75.10% in 10 places, which is better than other comparison methods. Finally, to adapt to the regulatory needs of the power system, the probability interval predictions are carried out based on the corrected wind speed by the proposed method. The probability interval coverage probability reaches 0.924 ~ 0.937, while the probability interval averaged width reaches 1.869 ~ 2.198 in 10 places. [ABSTRACT FROM AUTHOR]

Surface air temperature is a comprehensive function of aerosols in the atmosphere and various weather factors. However, there is no real-time aerosol concentration feedback in most operational numerical weather prediction (NWP) models. This raises a scientific question of how abnormal changes in air pollutants in a short period of time will affect the temperature prediction skill of NWP models. Thus, the study was carried out to investigate the possible influence of air pollution on the temperature forecast skill based on the operational NWP model over the Beijing–Tianjin–Hebei (BTH) region during January–February 2020. The results show that the average concentrations of PM2.5, SO2, NO2 and CO over the BTH region in February were smaller than those in January by 38.5%, 35.1%, 48.0% and 33.1%, respectively. Simultaneously, the forecast skills for surface temperature in February from both regional (RMAPS, Rapid-refresh Multi-scale Analysis and Prediction System) and global (ECMWF, European Centre for Medium-Range Weather Forecasts) operational NWP models improved markedly compared with that in January. In both models, the underestimation of maximum temperature and the overestimation of minimum temperature in most cities over the BTH region in February were significantly reduced. With the 24 h (24 h) forecast lead time, the RMSE (root mean square error) of BTH daily mean, maximum and minimum temperature prediction in February based on RMAPS were 17.3%, 9.8% and 21.6% lower than that in January, respectively. These are generally consistent with the other statistical indices such as deviation and regression coefficient. As the forecast lead time extended to 48 h and 72 h forecast, the phenomena still existed and were also evident in the ECMWF model. The improvement of temperature forecast skill of NWP models may be attributed to the unexpected dramatical reduction of air pollutants. Less aerosols during the daytime allow more solar radiation reaching the surface and cause a warming in the near-surface temperature, while less aerosols during the nighttime favor the outgoing long-wave radiation and then lead to a cooling near the ground. [ABSTRACT FROM AUTHOR]

The diurnal water‐vapour cycle is a critical component of the hydrological cycle, yet it is one of the hardest components to accurately reproduce in forecast and climate models. Previous studies have shown that both forecast and climate models underrepresent the diurnal water‐vapour cycle, which leads to errors in precipitation, cloud, and radiative transfer parameters. Most diurnal cycle studies were conducted in the Tropics, and very few model evaluations of this process exist for the Arctic. Additionally, the majority of studies focus on total column water‐vapour cycles; almost none use height‐resolved measurements. In this study, we evaluate the diurnal water‐vapour cycles in Environment and Climate Change Canada's Global Environmental Multiscale–High Resolution Deterministic Prediction System (GEM–HRDPS) numerical weather forecast model and the European Centre for Medium‐Range Weather Forecasts Reanalysis v5 (ERA5) using a Vaisala preproduction differential absorption lidar (DIAL) and a co‐located Global Positioning System (GPS) located in Iqaluit, Nunavut (63.75° N, 68.55° W). Both numerical products reproduce the phase of the diurnal cycle well below 1 km year‐round. However, ERA5's diurnal amplitudes are significantly smaller than the DIAL and GPS amplitudes. GEM–HRDPS amplitudes are consistently larger in the first few hundred metres but smaller above 1 km; it also underrepresents the amplitude of the total column diurnal cycle. Neither ERA5 nor GEM–HRDPS accurately reproduce the 12‐hr component of the diurnal cycle in either the height‐resolved or total column cycles. The inability to accurately reproduce the 12‐hr component in both numerical products suggests that the representation of some underlying process is incomplete, which can impact the accuracy of precipitation and radiative transfer algorithms. In conclusion, we find that the numerical products are able to reproduce the general behaviour and shape of the cycle but still require improvement at certain altitudes. [ABSTRACT FROM AUTHOR]

Transport in numerical weather models is typically restricted to advection by the resolved wind field, and representation of other flow processes relies on forcing or coupling to other models. A discussion of non‐advective transport in numerical weather models is presented and illustrated using two examples. These are the sea breeze and surface water flow. Simple models of these phenomena are represented in an adapted version of the Met Office Unified Model (UM) by means of modifying the existing cold‐pool scheme in the UM. This approach of incorporating different transport processes in one framework could facilitate a better representation of the Earth system, by increasing interaction between meteorological, surface and subsurface processes. [ABSTRACT FROM AUTHOR]

Nowadays, several global ensembles (GEs) which consist of several tens of members are being run operationally. In order to locally improve the probabilistic forecasts, various forecasting centers and research institutes utilize the GEs as initial and boundary conditions to drive regional convection permitting ensembles (RCPEs). RCPEs demand significant computer resources and often a limited number of ensemble members is affordable, which is smaller than the size of the driving GE. Since each RCPE member obtains the initial and boundary conditions from a specific GE member, there are many options to select the GE members. The study uses the European Centre for Medium‐Range Weather Forecasts (ECMWF) GE consisting of 50 members, to drive 20 members of COSMO model RCPE over the Eastern Mediterranean. We compare various approaches for automatic selection of the GE members and propose several optimal methods, including a random selection, which consistently lead to a better performance of the driven RCPE. The comparison includes verification of near surface variables and precipitation using various verification metrics. The results are validated using several methods of model physics perturbation. Besides the selection of the optimal ensemble configurations, we show that at high precipitation intensities spatial up‐scaling is recommended in order to obtain useful probabilistic forecasts. [ABSTRACT FROM AUTHOR]

The amount of solar irradiation that reaches Earth's surface is a key quantity of solar energy research and is difficult to predict, because it is directly affected by the changing constituents of the atmosphere. The numerical weather prediction (NWP) model performs computational simulations of the evolution of the entire atmosphere to forecast the future state of the atmosphere based on the current state. The Weather Research and Forecasting (WRF) model is a mesoscale NWP. WRF solar is an augmented feature of WRF, which has been improved and configured specifically for solar energy applications. The aim of this paper is to evaluate the performance of the high resolution WRF solar model and compare the results with the low resolution WRF solar and Global Forecasting System (GFS) models. We investigate the performance of WRF solar for a high-resolution spatial domain of resolution 1 × 1 km and compare the results with a 3 × 3 km domain and GFS. The results show error metrices rMAE {23.14%, 24.51%, 27.75%} and rRMSE {35.69%, 36.04%, 37.32%} for high resolution WRF solar, coarse domain WRF solar and GFS, respectively. This confirms that high resolution WRF solar performs better than coarse domain and in general. WRF solar demonstrates statistically significant improvement over GFS. [ABSTRACT FROM AUTHOR]

The Yin‐He Global Spectral Model (YHGSM) is a dry‐mass conserving hydrostatic global spectral model, relying on spectral transforms to compute horizontal derivatives. We present an extension of YHGSM core named YHGSM‐FVM which uses a second‐order finite‐volume method (FVM) to compute the horizontal derivatives in grid‐point space instead of the spectral approach. With this approach, the computational efficiency of the spectral model is improved since part of the spectral transforms is superseded by FVM which only needs local data and the computational demand is lower. More importantly, YHGSM‐FVM is still a spectral model solving the Helmholtz equation directly in spectral space with a highly efficient semi‐implicit semi‐Lagrangian advection scheme. The comparisons between YHGSM‐FVM and YHGSM are conducted, and the results show that both models have comparable prediction skill, but YHGSM‐FVM outperforms YHGSM in computational efficiency. [ABSTRACT FROM AUTHOR]

Precipitable water vapor (PWV) monitoring with real-time precise point positioning (PPP) is required for the improved early detection of increasingly common extreme weather occurrences. This study takes Hong Kong as the research object. The aim is to explore the accuracy of real-time global navigation satellite system (GNSS) PPP in estimating PWV at low latitudes and under different weather conditions. In this paper, real-time PPP is realized by using observation data from continuously operating reference stations (CORS) in Hong Kong and real-time products from the Centre National d'Etudes Spatiales (CNES). The Tm model calculated using numerical weather prediction (NWP) data converts the zenith tropospheric delay (ZTD) of real-time PPP inversion into PWV and evaluates its accuracy using postprocessing products. The experimental results show that compared with GPS, multi-GNSS can reduce the convergence time of PPP by 29.20% during rainfall periods and by 12.06% during nonrainfall periods. The improvement in positioning accuracy is not obvious, and the positioning accuracy of the two is equivalent. Real-time PPP ZTD experiments show that there are lower average values for bias, standard deviation (STDEV), and root mean square (RMS) during nonrainfall periods than during rainfall periods. Real-time PPP PWV experiments show that there are also lower bias, STDEV, and RMS values during nonrainfall periods than during rainfall periods. The comparative study between rainfall and nonrainfall periods is of great significance for the real-time monitoring and forecasting of water vapor changes. [ABSTRACT FROM AUTHOR]

Cloud microphysics parametrizations control the transfer of water between phases and hydrometeor species in numerical weather prediction and climate models. As a fundamental component of weather modelling systems cloud microphysics can determine the intensity and timing of precipitation, the extent and longevity of cloud cover and its impact on radiative balance, and directly influence near surface weather metrics such as temperature and wind. In this paper we introduce and demonstrate the performance of a double moment cloud microphysical scheme (CASIM: Cloud AeroSol Interacting Microphysics) in both midlatitude and tropical settings using the same model configuration. Comparisons are made against a control configuration using the current operational single moment cloud microphysics, and CASIM configurations that use fixed in‐cloud droplet number or compute cloud droplet number concentration from the aerosol environment. We demonstrate that configuring CASIM as a single moment scheme results in precipitation rate histograms that match the operational single moment microphysics. In the midlatitude setting, results indicate that CASIM performs as well as the single moment microphysics configuration, but improves certain aspects of the surface precipitation field such as greater extent of light (<$$ < $$1 mm ·$$ \cdotp $$ hr−1$$ {}^{-1} $$) rain around frontal precipitation features. In the tropical setting, CASIM outperforms the single moment cloud microphysics as evident from improved comparison with radar derived precipitation rates. [ABSTRACT FROM AUTHOR]

Local ensemble transform Kalman filters (LETKFs) allow explicit calculation of the Kalman gain, and by this the contribution of individual observations to the analysis field. Though this is a known feature, the information on the analysis contribution of individual observations (partial analysis increment) has not been used as systematic diagnostic up to now despite providing valuable information. In this study, we demonstrate three potential applications based on partial analysis increments in the regional modelling system of Deutscher Wetterdienst and propose their use for optimising LETKF data assimilation systems, in particular with respect to satellite data assimilation and localisation. While exact calculation of partial analysis increments would require saving the large, five‐dimensional ensemble weight matrix in the analysis step, it is possible to compute an approximation from standard LETKF output. We calculate the Kalman gain based on ensemble analysis perturbations, which is an approximation in the case of localisation. However, this only introduces minor errors, as the localisation function changes very gradually among nearby grid points. On the other hand, the influence of observations always depends on the presence of other observations and settings for the observation error and for localisation. However, the influence of observations behaves approximately linearly, meaning that the assimilation of other observations primarily decreases the magnitude of the influence, but it does not change the overall structure of the partial analysis increments. This means that the calculation of partial analysis increments can be used as an efficient diagnostic to investigate the three‐dimensional influence of observations in the assimilation system. Furthermore, the diagnostic can be used to detect whether the influence of additional experimental observations is in accordance with other observations without conducting computationally expensive single‐observation experiments. Last but not least, the calculation can be used to approximate the influence an observation would have when applying different assimilation settings. [ABSTRACT FROM AUTHOR]

Convection‐permitting (CP) models have provided a step change in the ability to forecast impactful convective storms, which pose risks such as flash flooding and lightning. Despite CP models being routinely run over tropical Africa, they typically lack data assimilation and are initialised directly from a global analysis. The CP model therefore takes time to generate convective structures and consequently precipitation, rendering the early part of the forecast unusable. A 'warm‐starting' method developed within the Met Office Unified Model is tested over a large tropical Africa domain, leading to substantial improvements in rainfall predictions during both the spin‐up period but also later in the forecast. The spin‐up period is found to vary considerably, both geographically and with model initialisation time, due to the differing states of the atmosphere through the day. We quantify this impact through a spin‐up diagnostic, which shows that over central Africa the spin‐up period can be as large as 24 hr. In these regions, the warm starting substantially reduces spin‐up time. Additionally, the large tropical Africa domain allows us to examine the impact of the prevailing meteorology on both the spin‐up period and the warm‐start method. We find large‐scale positive vorticity filaments over West Africa, likely tied to African Easterly Waves, increase rainfall rates across all models. The substantial reduction in the spin‐up of rainfall and increased skill at longer lead times in the warm‐start demonstrate the importance of representing convective features in the analysis for a CP model. This is especially important in large tropical domains where the initial conditions are likely to have a lasting impact on the forecast. Improvements in model performance across the forecast will enhance the ability of forecasters to provide advice on high‐impact weather events, supporting increased resilience against weather extremes in Africa. [ABSTRACT FROM AUTHOR]

Forecasts of numerical weather prediction models unavoidably contain errors, and it is a common practice to post-process the model output and correct the error for the proper use of the forecasts. This study develops a grid-to-multipoint (G2N) model output error correction scheme which extracts model spatial features and corrects multistation forecasts simultaneously. The model was tested for an operational high-resolution model system, the precision rapid update forecasting system (PRUFS) model, running for East China at 3 km grid intervals. The variables studied include 2 m temperature, 2 m relative humidity, and 10 m wind speed at 311 standard ground-based weather stations. The dataset for training G2N is a year of historical PRUFS model outputs and the surface observations of the same period and the assessment of the G2N performance are based on the output of two months of real-time G2N. The verification of the real-time results shows that G2N reduced RMSEs of the 2 m temperature, 2 m relative humidity, and 10 m wind speed forecast errors of the PRUFS model by 19%, 24%, and 42%, respectively. Sensitivity analysis reveals that increasing the number of the target stations for simultaneous correction helps to improve the model performance and reduces the computational cost as well indicating that enhancing the loss function with spatial regional meteorological structure is helpful. On the other hand, adequately selecting the size of influencing grid areas of the model input is also important for G2N to incorporate enough spatial features of model forecasts but not to include the information from the grids far from the correcting areas. G2N is a highly efficient and effective tool that can be readily implemented for real-time regional NWP models. [ABSTRACT FROM AUTHOR]

GPS PWV (Precipitable Water Vapour) is estimated using three years ECMWF, JRA55, MERRA2, and NCEP datasets for two IGS stations HYDE and IISC in India to assess their accuracies over the Indian region. The NWP derived temperature, pressure, zenith hydrostatic delay (ZHD), zenith wet delay (ZWD), and PWV values are validated using AWS observations and derived products. The obtained PWV values are also compared with the total column water product from ECMWF. The results show absolute mean temperature differences of 1-2 °K and pressure differences of 1-4 hPa for HYDE and IISC. NWP models' pressure and temperature values strongly correlate with AWS values with R² close to 0.9 in most of the cases. The estimated zenith tropospheric delay (ZTD) shows an average difference between 4mm and 9mm with respect to IGS ZTD. The mean differences between NWP and reference PWV values are observed from 1mm to 3 mm. All the derived products obtained using NWP datasets are close to the reference AWS products. However, PWV derived using ECMWF, JRA55 and NCEP is very much close to AWS PWV and can be used for precise PWV computation in this region. [ABSTRACT FROM AUTHOR]

This is a comprehensive approach for the analysis of fog attributes with high-density intra-hourly data (Oct–Mar) for a period of 11 years from Jan 2009 to Dec 2019 in sub-urban Bengaluru. Also, the characteristics of nocturnal inversion parameters have been examined on the days with/without fog to explore any relationship between the two concurring meteorological phenomena. Lastly, ability of NWP model (NCMRWF, Unified Model, NCUM) was examined for predicting fog over Kempegowda International Airport, Bengaluru. Under the studied time frame, December reported 129 and January 126 fog events, while October, November, February, and March reported 33, 96, 44, 11 events, respectively. Similarly, December and January accounted for maximum fog hours. More than 80% of fog events were reported between 2200 and 0200 UTC with maximum (74) in the 0100–0129 UTC slot and dispersed between 0000 and 0400 UTC. Intensity-wise, shallow fog is most frequent (51%), while moderate, dense, and very-dense have a frequency of 29, 19.6 and 0.4%, respectively. As for atmospheric inversions in Bengaluru, elevated inversions are common occurrences throughout the year, while surface inversions are more frequent in the winter. In the study, it was found that 35% fog events coincided with two elevated inversions (EI) under the 700 hPa ceiling, whereas 25% occurred with a single EI. Surface inversions coinciding with fog events are relatively rare and mostly (71%) occur alongside EI. NWP model predicted reduced visibility during fog events with a hit rate of 42%, false alarm rate of 5% and bias value of 0.76. Model produced mixed results in forecasting of meteorological parameters and visibility. Wind speed and visibility were underestimated, whereas temperature and relative humidity show good agreement with the observed values. [ABSTRACT FROM AUTHOR]

The multi-model ensemble (MME) forecast for meteorological elements has been proved many times to be more skillful than the single model. It improves the forecast quality by integrating multiple sets of numerical forecast results with different spatial-temporal characteristics. Currently, the main numerical forecast results present a grid structure formed by longitude and latitude lines in space and a special two-dimensional time structure in time, namely the initial time and the lead time, compared with the traditional one-dimensional time. These characteristics mean that many MME methods have limitations in further improving forecast quality. Focusing on this problem, we propose a deep MME forecast method that suits the special structure. At spatial level, our model uses window self-attention and shifted window attention to aggregate information. At temporal level, we propose a recurrent like neural network with rolling structure (Roll-RLNN) which is more suitable for two-dimensional time structure that widely exists in the institutions of numerical weather prediction (NWP) with running service. In this paper, we test the MME forecast for sea level pressure as the forecast characteristics of the essential meteorological element vary clearly across institutions, and the results show that our model structure is effective and can make significant forecast improvements. [ABSTRACT FROM AUTHOR]

Two‐fluid modelling has recently emerged as a promising approach to representing cumulus convection in weather and climate models. This study applies the two‐fluid model described in Part II to a shallow cumulus convection case study over land (ARM). Large‐eddy simulation data are used to tune the majority of the closures that determine the properties of entrained and detrained air. The two‐fluid model is generally able to reproduce the profiles of the mean and turbulent quantities over all stages of the diurnal cycle. As such, the initiation of shallow convection and the evolution of the cloud layer are well captured. The robustness of the two‐fluid model is verified further using a steady‐state test case (BOMEX), in which the cloud properties are also well modelled. [ABSTRACT FROM AUTHOR]

Observation system experiments (OSEs) are conducted to assess the potential impacts of horizontal line‐of‐sight wind profile observations from the Aeolus satellite on tropical cyclone (TC) forecasting. The OSEs utilize the operational Hurricane Weather and Research Forecasting (HWRF) model. The OSEs include 226 forecasts from seven TC cases in the Atlantic and Eastern Pacific basins. Comparisons between Aeolus and model background winds show that winds from Aeolus are consistently stronger than those from HWRF. Data assimilation statistics also demonstrate that the greatest potential impacts from the assimilation of Aeolus observations are likely to occur in the upper troposphere and lower stratosphere and within approximately 500 km from the TC centre. For TC forecasting applications, the assimilation of Aeolus observations improves TC intensity and size forecasts in the Eastern Pacific basin, while the results for track forecasts and results from the Atlantic basin are mixed. However, in both basins, the largest and most statistically significant, positive impacts from the assimilation of Aeolus observations occur when reconnaissance flight data are unavailable and during the early stages of TC development. The traditionally used forecast assessments of TC intensity, track and size are rooted in surface‐based metrics, and an additional investigation above the surface demonstrated larger improvements from assimilating Aeolus observations on TC wind structure above 400 hPa as compared to the lower troposphere. Several, different assessments throughout this study demonstrate higher uncertainty and the need for special consideration associated with assimilation techniques within 500 km from the TC centre. [ABSTRACT FROM AUTHOR]

A new method based on the Transformer model is proposed for post‐processing of numerical weather prediction (NWP) forecasts of 2 m air temperature. The Transformer is a machine learning (ML) model based on self‐attention, which extracts information about which inputs are most important for the prediction. It is trained using time series input from NWP variables and crowd‐sourced 2 m air temperature observations from more than 1000 private weather stations (PWSs). The performance of the new post‐processing model is evaluated using both observational data from PWSs and completely independent observations from the Danish Meteorological Institute (DMI) network of surface synoptic observations (SYNOP) stations. The performance of the Transformer model is compared against the raw NWP forecast, as well as against two benchmark post‐processing models; a linear regression (LR) model and a neural network (NN). The results evaluated using PWS observations show an improvement in the 2 m temperature forecasts with respect to both bias and standard deviation (STD) for all three post‐processing models, with the Transformer model showing the largest improvement. The raw NWP forecast, LR, NN and Transformer model have a bias and STD of 0.34 and 1.96°C, 0.03 and 1.63°C, 0.10 and 1.53°C and 0.02 and 1.13°C, respectively. The corresponding results using DMI SYNOP stations also show improved forecasts, where the Transformer model performs better than both the raw NWP forecast and the two benchmark models. However, a dependence on distance to the coast and cold temperatures is observed. [ABSTRACT FROM AUTHOR]

The Met Office held a testbed over winter 2020/2021 where a new numerical weather prediction (NWP) sub‐km ensemble was set up on‐demand in response to interesting weather phenomena in the United Kingdom. The domain for the model was chosen in real time by a community of Met Office Research Scientists and Operational Meteorologists and over a 4‐month period the ensemble was triggered for nine events. The purpose of the testbed was to investigate whether a real‐time weather regime‐based enhancement in NWP capability was feasible, to understand what benefits a testbed environment might give, and to explore the practicalities of running such an event. Case studies from the testbed demonstrated that forecast ensembles at 2.2 km and 300 m grid spacing were able to capture observed winter weather, with greater spatial detail apparent, especially over complex orography, in the 300‐m model. Ensemble spread appeared less influenced by resolution, potentially due to the size of the domains tested or the weather regimes of the case studies. The testbed also showcased underutilized observations and additional radiosonde ascents were conducted. All the testbed meetings were conducted virtually due to COVID‐19 restrictions, and decisions were made about when to trigger the event using an online message board. The winter 2020/2021 testbed provides ideas for how on‐demand weather‐dependent testbeds might be conducted in the future. However, several recommendations are made that would enhance testbed benefits further, including more dedicated resource, stronger technology and data visualization and greater participation from both academia and weather information users. [ABSTRACT FROM AUTHOR]

Data impact experiments are conducted employing the Polar Weather Research and Forecasting (WRF) model during the YOPP‐SH summer special observing period (SOP) using the Antarctic Mesoscale Prediction System (AMPS) framework to determine the forecast impact of numerous additional radiosondes collected during the SOP. Hybrid variational‐ensemble three‐dimensional data assimilation is performed on model forecast domains over Antarctica and the Southern Ocean using all regular observations normally available (Experiment "NoSOP") and using the same set plus the extra soundings launched for the SOP (Experiment "SOP"). The SOP results show better near‐surface temperature and wind‐speed forecasts than the NoSOP results, primarily over West Antarctica. Radiosonde profiles confirm that temperature and wind‐speed forecasts are improved throughout the troposphere with the addition of the SOP radiosonde data, but the results for relative humidity are variable. Temperatures are improved at lower levels early in the forecasts, whereas wind speeds are better at higher levels later in the forecasts. An evaluation against the ERA5 global reanalysis that provides a much broader perspective reveals that the improved forecast skill for the SOP experiment persists up to 72 hours for temperature, wind speed, and relative humidity. The gains, however, are primarily confined to the Antarctic continent, consistent with the additional radiosonde spatial coverage being mainly poleward of 60°S. With extra radiosondes concentrated over the Antarctic Peninsula, SOP forecasts of the region downstream of the Peninsula were significantly improved compared to NoSOP forecasts. In addition, it is found that the assimilation of the additional radiosonde data can have a greater impact on the forecasts of strong cyclones, as shown for a major coastal cyclone affecting West Antarctica, with improvements in its magnitude and track. The results also suggest that increasing radiosonde launches at lower southern latitudes would improve forecasts over the Southern Ocean, especially during austral winter. [ABSTRACT FROM AUTHOR]

In spite of WRF-Solar being a numerical weather prediction (NWP) model specifically tailored for solar energy applications, the model lacks a specific cloud initialization component to improve short-range predictions. To overcome this limitation we have combined the fundamental concepts of the Multi-sensor Advection Diffusion nowCast (MADCast) model of using satellite observations to detect the clouds and advect and diffuse the clouds within a NWP model, with the better cloud–aerosol–radiation physics of WRF-Solar, into a new model referred to as MAD-WRF. The MAD-WRF cloud initialization combines a cloud parameterization that infers the presence of clouds based on relative humidity with observations of the cloud mask and cloud top/base height to provide a three-dimensional cloud analysis (liquid, ice, and snow hydrometeor content). During the forecasts, the hydrometeors can be advected and diffused with no microphysics, in what we refer to as the MAD-WRF passive mode. Alternatively, these passive hydrometeors can be integrated into the explicitly resolved hydrometeors during a nudging phase, designated the MAD-WRF active mode. In the nudging phase, there is a relaxation of the resolved hydrometeors towards the passive hydrometeors. Both modes aim to avoid an abrupt dissipation of clouds if the atmospheric environment is not adequate to support them. Both MAD-WRF active and passive modes as well as WRF-Solar have been run in a demonstration of MAD-WRF over the contiguous U.S., and results from these short-range forecasts (0–6 h) are presented here to illustrate the added value of MAD-WRF for global horizontal irradiance predictions. These results clearly illustrate the added value that MAD-WRF brings for short term irradiance predictions with the WRF-Solar model. • We have created the MAD-WRF model to improve cloud initialization in WRF-Solar. • MAD-WRF improves WRF-Solar short range irradiance predictions. • Consistent added value with respect to climatology. [ABSTRACT FROM AUTHOR]

Granularity of the grid (both horizontally and vertically) is a key consideration when conducting localised Numerical Weather Prediction (NWP) modelling. Generally speaking, an NWP model with a finer grid can explicitly resolve more processes and require less parameterisation. However, a finer grid also requires more computation and it is not always clear that a finer grid will produce a more accurate forecast. In this study, we explore the sensitivity of rainfall prediction over Singapore to grid resolution. We use the Weather and Research Forecasting model (WRF) to forecast rainfall over Singapore and explore the performance of horizontal resolutions ranging from 1 km to 12 km. We test the performance on a set of dates from across the years 2020–2021 against both ground observations and radar-derived rain rates. When compared to ground observations, we show that, overall, the higher resolution produces the highest Critical Success Index (CSI) for rain rates in excess of 0.5 mm/h. When compared against radar-derived rain rates, the 1 km domain produces superior CSI values for all rain rates. The daily-average hourly Fractional Skill Score (FSS) was then calculated for some dates and showed agreement with the CSI results with the exception of a north-east monsoon day where, for heavier rain rates, the 3 km domain has superior FSS. We also investigate a particularly heavy rain event on 10 January 2021 and show that the 3 km grid has highest CSI for rain rates of 4 mm/h and 16 mm/h (based on both ground-based and radar-derived rain rates), however, the 1 km has superior FSS for this event. [ABSTRACT FROM AUTHOR]

The Tropical Cyclone (TC) track prediction using different NWP models and its verification is the critical task to provide prior knowledge about the model errors, which is beneficial for giving the model guidance-based real-time cyclone warning advisories. This study has attempted to verify the Global Forecast System (GFS) model forecasted tropical cyclone track and intensity over the North Indian Ocean (NIO) for the years 2019 and 2020. GFS is one of the operational models in the India Meteorological Department (IMD), which provides the medium-range weather forecast up to 10 days. The forecasted tracks from the GFS forecast are obtained using a vortex tracker developed by Geophysical Fluid Dynamics Laboratory (GFDL). A total of 13 tropical cyclones formed over the North Indian Ocean, eight during 2019 and five in 2020 have been considered in this study. The accuracy of the model predicted tracks and intensity is verified for five days forecasts (120 h) at 6-h intervals; the track errors are verified in terms of Direct Position Error (DPE), Along Track Error (ATE) and Cross-Track Error (CTE). The annual mean DPE over NIO during 2019 (51-331 km) is lower than 2020 (82-359 km), and the DPE is less than 150 km up to 66 h during 2019 and 48 h during 2020. The positive ATE (76-332 km) indicates the predicted track movement is faster than the observed track during both years. The positive CTE values for most forecast lead times suggest that the predicted track is towards the right side of the observed track during both years. The cyclone Intensity forecast for the maximum sustained wind speed (MaxWS) and central mean sea level pressure (MSLP) are verified in terms of mean error (ME) and root mean square error (RMSE). The errors are lead time independent. However, most of the time model under-predicted the cyclone intensity during both years. Finally, there is a significant variance in track and intensity errors from the cyclone to cyclone and Bay of Bengal basin to the Arabian Sea basin. [ABSTRACT FROM AUTHOR]

Wind information in urban areas is essential for many applications related to air pollution, urban climate and planning, safety of drone‐related operations, and assessment of urban wind energy potential. These applications require accurate wind forecasts, and obtaining this information in an urban environment is challenging as the morphology of a city varies from street to street, altering the wind flow. Remote sensing techniques such as Doppler lidars (light detection and ranging) provide a unique opportunity for wind forecast verification as they can provide both the vertical profile of the horizontal wind and the spatial variation in the horizontal domain at high resolution. In this study, the performance of numerical weather prediction (NWP) models, analysis systems, and large‐eddy simulation (LES) models have been analysed by comparing the modelled winds against Doppler lidar observations under various atmospheric conditions and from season to season, in the coastal environment of Helsinki, Finland. The long‐term mean vertical profile of the modelled horizontal wind shows good agreement with observations; the NWP model and the analysis systems selected here exhibit different strengths and weaknesses depending on the atmospheric conditions but no significant diurnal variation in performance. However, both the model and analysis systems show differences in their spatially‐averaged bias when investigating different wind directions. LES verification shows that these models can potentially provide winds down to street level, given pre‐computed scenarios of atmospheric conditions. For Helsinki, the observed winds are stronger during winter than summer, and, on average, higher wind speeds were observed at the urban site than the sub‐urban site. [ABSTRACT FROM AUTHOR]

Developments in the assimilation of satellite data in numerical weather prediction (NWP), from the first experiments in the late 1960s to the present day, are presented in a two-part review article. This part, Part II, reviews the progress in recent years, from about 2000. It includes summaries of advances in the relevant satellite remote-sensing technologies and in methods to assimilate observations from these instruments into NWP systems. It also summarises impacts on forecast skill. Continued progress has been made on the assimilation of passive infrared (IR) sounding data and microwave (MW) sounding and imaging data. This has included data from hyperspectral IR sounders, which first became available during this period. Advances in the use of cloud-affected radiances, from both IR and MW instruments, have been made. In support of this progress, further developments have been made in fast radiative transfer models and in bias correction techniques, and work has continued to improve understanding and representation of observation uncertainties. Continued progress has also been made on the use of wind information from satellites, including atmospheric motion vectors and scatterometer data. A new source of temperature and humidity information, from radio occultation observations, has become available during the period and has been exploited by many NWP centres. The impact of satellite data on NWP accuracy is continually assessed using a range of methods and metrics. Some results from recent Observing System Experiments (OSEs) and Forecast Sensitivity to Observation Impact (FSOI) assessment are presented and other methods are discussed. The role of satellite data in NWP-based atmospheric reanalysis systems is also described. [ABSTRACT FROM AUTHOR]

• Direct normal irradiance (DNI) is computed by a physics-based solution of the gross solar radiation in the circumsolar region. • The new DNI model is coupled with a numerical weather prediction model to forecast day-ahead DNI over the contiguous United States. • The significant bias in the conventional forecasting models has been remarkably reduced. Direct normal irradiance (DNI) is often interpreted differently in ground measurements and forecasts using numerical weather prediction (NWP) models, leading to substantial bias in DNI forecasting especially under cloudy-sky conditions. To mitigate the bias, we use the Fast All-sky Radiation Model for Solar applications with DNI (FARMS-DNI) to conduct physics-based simulations of solar radiation in the circumsolar region. The DNI is quantified using the sum of the radiation along the sun direction and the scattered radiation within the circumsolar region. FARMS-DNI is coupled with the Weather Research and Forecasting model with solar extensions (WRF-Solar) to predict day-ahead DNI in 9-km pixels over an extended domain covering the contiguous United States. The DNI forecasts for the entire year of 2018 are validated using surface-based observations at the Atmospheric Radiation Measurement (ARM) and Surface Radiation Budget Network (SURFRAD) sites and a satellite-based solar radiation product developed by the procedure of the National Solar Radiation Data Base (NSRDB). The results show that the WRF-Solar coupled with the conventional Beer-Bouguer-Lambert law underestimes the overall DNI forecasts by 30%-60%, which is a joint effect of the uncertainties in computing the circumsolar radiation and forecasting cloud properties. This bias is significantly reduced by using FARMS-DNI. [ABSTRACT FROM AUTHOR]

Comprehending the mechanisms underlying the response of water productivity to changing climate is essential for developing effective strategies that can maintain sustainable agricultural productivity and ecological stability amidst water scarcity and climate change. This study aimed to clarify the effects of climate and associated biologic and abiotic factors on the net water productivity (NWP) in the winter wheat—summer maize rotation system within the North China Plain (NCP), through integrating spatial-temporal pattern analysis, the Optimal Parameter-based Geographical Detector model, Random Forest, and Structural Equation Modeling. Despite climate warming and water scarcity, the NWP in the vast majority of rotation systems within the NCP exhibited an increasing trend between 2000 and 2019, which was anticipated to persist in the future. From 2000–2019, the impact of climate on the NWP exerted a fluctuating diminishment, besides the influence of WS. It is essential to add adaptive strategies that can cope with WS fluctuation in future management, to better tackle climate variability. Apart from direct effects, meteorological factors indirectly induced NWP variation primarily by influencing edaphic moisture dynamic and crop growth factors. In the paths underlying the effect of climate change on NWP, positive and negative effects coexisted. Edaphic moisture within the 40–100 cm layer played an important role in affecting NWP throughout rotation systems. To sustainably improve NWP, it was advisable to adopt deeper-rooting crop varieties in conjunction with moderate deficit irrigation and improving topsoil water retention capability. Overall, we incorporated soil and crop factors into the framework of climate impact research, providing a systematic elucidation for the mechanisms through which climate change affects NWP. Additionally, the conclusions were based on a continuous spatiotemporal framework that accounted for the geospatial heterogeneity of variables, thereby giving conclusions ample regional universality. • Soil & crop factors can help clarify mechanisms underlying climate's effect on NWP. • Despite warming and water scarcity, NWP increased and is expected to persist in future. • Impact of climate variability on NWP in wheat-maize rotation is diminishing. • Soil moisture at 40–100 cm depth is important for sustaining NWP. [ABSTRACT FROM AUTHOR]

This paper proposes a simple and powerful optimal integration (OPI) method for improving hourly quantitative precipitation forecasts (QPFs, 0-24 h) of a single-model by integrating the benefits of different biascorrected methods using the high-resolution CMA-GD model from the Guangzhou Institute of Tropical and Marine Meteorology of China Meteorological Administration (CMA). Three techniques are used to generate multi-method calibrated members for OPI: deep neural network (DNN), frequency-matching (FM), and optimal threat score (OTS). The results are as follows: (1) The QPF using DNN follows the basic physical patterns of CMA-GD. Despite providing superior improvements for clear-rainy and weak precipitation, DNN cannot improve the predictions for severe precipitation, while OTS can significantly strengthen these predictions. As a result, DNN and OTS are the optimal members to be incorporated into OPI. (2) Our new approach achieves state-of-the-art performances on a single model for all magnitudes of precipitation. Compared with the CMA-GD, OPI improves the TS by 2.5%, 5.4%, 7.8%, 8.3%, and 6.1% for QPFs from clear-rainy to rainstorms in the verification dataset. Moreover, OPI shows good stability in the test dataset. (3) It is also noted that the rainstorm pattern of OPI relies heavily on the original model and that OPI cannot correct for deviations in the location of severe precipitation. Therefore, improvements in predicting severe precipitation using this method should be further realized by improving the numerical model's forecasting capability. [ABSTRACT FROM AUTHOR]

Extreme rainfall events have posed several serious threats to many populated and urbanized areas in the world including the Indian subcontinent. Therefore, accurate predictions of their intensity and areas of influence are important for flood‐prone risk assessments. On 14 July 2016, heavy to exceptionally heavy rainfall occurred in Yamunanagar (30.16° N, 77.29° E), located in the state of Haryana in North India, which led to widespread disruption of communication, electricity, inundation of houses, and so forth. The present paper aims at examining observational, synoptic, thermo‐dynamical, and numerical features associated with this devastating rainfall episode. The analysis found that during extreme rainfall episodes, a trough in mid‐tropospheric westerlies and a strong low‐level atmospheric monsoonal flow seem to have strongly interacted with each other, creating a strong convergence zone near study areas that led to a severe rainstorm. The quasi‐stationary supercells were also noticed due to continuous moisture incursions from the Bay of Bengal and orographic uplift over the Himalayas near Yamunanagar. A deep layer of wind shear interacts dynamically with the convergence zone and leads to a potential rainstorm. Thermodynamic indices indicate high instability over the heavy rainfall area. The dynamics of this event were studied in detail by using three‐dimensional variational data assimilation within the weather research and forecasting model, configured with triple two‐way nesting domains (27, 9, and 3 km). The model results show that the weather research and forecasting model satisfactorily captures the quantitative precipitation (300 mm) in 24 h over the Yamunanagar region as compared with observation (365 mm). [ABSTRACT FROM AUTHOR]

In the framework of the EUMETSAT Polar System–Second Generation (EPS‐SG) preparation, a new generation of the Infrared Atmospheric Sounding Interferometer (IASI) instrument has been designed. The IASI‐New Generation (IASI‐NG) will measure radiances at a doubled spectral resolution compared to its predecessor and with a signal‐to‐noise ratio improved by a factor of 2. The large amount of data arising from IASI‐NG will present many challenges for data transmission, storage and assimilation. Moreover, the full set of measured radiances will not be exploitable in an operational numerical weather prediction (NWP) context. For these reasons, an appropriate IASI‐NG channel selection in needed, aiming to select the most informative channels for NWP models. Therefore, the standard iterative channel selection methodology, based on the optimal linear estimation theory and assuming spectrally correlated errors, has been applied to a set of simulated data of the IASI‐NG spectrum. The entire simulated IASI‐NG spectrum has been first investigated, while finally focusing the channel selection procedure on the most interesting wavelength ranges for the assimilation. Through this process, a total of 500 channels have been chosen to serve as a basis for future channel selections to be provided to NWP centres – 277 temperature, 23 surface‐sensitive and 200 water vapour channels. One‐dimensional variational (1D‐Var) assimilation experiments show that using this selected set of channels leads to a reduction of the standard deviation of the error in temperature (up to 30%) and water vapour (up to 50%) profiles with respect to the apriori information. [ABSTRACT FROM AUTHOR]

The philosophy of forecast verification is rather different between deterministic and probabilistic verification metrics: generally speaking, deterministic metrics measure differences, whereas probabilistic metrics assess reliability and sharpness of predictive distributions. This article considers the root‐mean‐square error (RMSE), which can be seen as a deterministic metric, and the probabilistic metric Continuous Ranked Probability Score (CRPS), and demonstrates that under certain conditions, the CRPS can be mathematically expressed in terms of the RMSE when these metrics are aggregated. One of the required conditions is the normality of distributions. The other condition is that, while the forecast ensemble need not be calibrated, any bias or over/underdispersion cannot depend on the forecast distribution itself. Under these conditions, the CRPS is a fraction of the RMSE, and this fraction depends only on the heteroscedasticity of the ensemble spread and the measures of calibration. The derived CRPS–RMSE relationship for the case of perfect ensemble reliability is tested on simulations of idealised two‐dimensional barotropic turbulence. Results suggest that the relationship holds approximately despite the normality condition not being met. [ABSTRACT FROM AUTHOR]

Two methods are widely used to assess the impact of observations in global numerical weather prediction (NWP): data denial experiments (DDEs) and the forecast sensitivity‐based observation impact (FSOI) method. A DDE measures the impact on forecast accuracy of removing an observation type from the system, whereas FSOI measures the amount by which an observation type reduces the short‐range forecast error (within a system containing all observation types). It is usually found, in global NWP experiments in which both methods are used, that the FSOI impact of a given observation type is greater than its DDE impact. The aim of this article is to present the theory behind the DDE and FSOI metrics that are commonly used and to explore factors that cause DDE and FSOI metrics to give different results. The general theory is presented for an optimal analysis and forecast system and then applied to a simple model with two state variables. To explain the commonly found result of FSOI impact greater than DDE impact, the following system properties are found to be important: mixing of information (and error) by the forecast model between state variables, the rate of forecast error growth, the presence of forecast model error, the way in which observational information is distributed between state variables and denied from them, and the presence of error in the data used for forecast verification. These results provide insight into why NWP systems are resilient to the removal of observational information; they are shown to be resilient if information is denied primarily from well‐observed variables but not when denied primarily from poorly observed variables. [ABSTRACT FROM AUTHOR]

We present an ensemble prediction system using a Deep Learning Weather Prediction (DLWP) model that recursively predicts six key atmospheric variables with six‐hour time resolution. This computationally efficient model uses convolutional neural networks (CNNs) on a cubed sphere grid to produce global forecasts. The trained model requires just three minutes on a single GPU to produce a 320‐member set of six‐week forecasts at 1.4° resolution. Ensemble spread is primarily produced by randomizing the CNN training process to create a set of 32 DLWP models with slightly different learned weights. Although our DLWP model does not forecast precipitation, it does forecast total column water vapor and gives a reasonable 4.5‐day deterministic forecast of Hurricane Irma. In addition to simulating mid‐latitude weather systems, it spontaneously generates tropical cyclones in a one‐year free‐running simulation. Averaged globally and over a two‐year test set, the ensemble mean RMSE retains skill relative to climatology beyond two‐weeks, with anomaly correlation coefficients remaining above 0.6 through six days. Our primary application is to subseasonal‐to‐seasonal (S2S) forecasting at lead times from two to six weeks. Current forecast systems have low skill in predicting one‐ or 2‐week‐average weather patterns at S2S time scales. The continuous ranked probability score (CRPS) and the ranked probability skill score (RPSS) show that the DLWP ensemble is only modestly inferior in performance to the European Center for Medium Range Weather Forecasts (ECMWF) S2S ensemble over land at lead times of 4 and 5–6 weeks. At shorter lead times, the ECMWF ensemble performs better than DLWP. Plain Language Summary: The world's leading weather forecasting institutions currently rely on computationally expensive weather models running on massive supercomputers. In order to have predictive skill for forecasts two to six weeks in the future, large ensembles of many nearly identical runs of these models are required, but the computational resources needed for these ensembles scales with the number of forecasts run. Since the resources needed rapidly approaches modern‐day computing limits, we explore the possibility of using computationally cheap weather models based on machine learning algorithms which learn to reproduce the evolution of weather. Our machine‐learning model is capable of running 320 forecasts in three minutes on a single workstation, while the state‐of‐the‐art model from the European Center for Medium‐Range Weather Forecasts (ECMWF) utilizes supercomputers to run 50 forecasts. Our ensemble weather model produces realistic forecasts of weather events such as Hurricane Irma in 2017 and is even capable of nearly matching the performance of the ECMWF ensemble for forecasts of temperature four to six weeks in the future. Key Points: An ensemble forecast system is developed using convolution neural networks (CNNs) to generate data‐driven global forecastsOnly 3 s are required to compute a large 320‐member ensemble of skillful 6‐week sub‐seasonal predictionsShorter lead time forecasts also show skill, including a single deterministic 4‐day forecast for Hurricane Irma [ABSTRACT FROM AUTHOR]

Numerical weather prediction (NWP) models, which attempt to simulate the full state of the atmosphere, come with many options for dealing with processes that are unable to be explicitly resolved by the model. These model parameterizations are an ongoing area of research in atmospheric science. However, with continuous contributions from the research community and subsequent upgrade of the NWP codes, there are often many options for each unresolved process--leaving the user confronted with potentially thousands of ways to configure the model. We use the weather research and forecasting (WRF) model to forecast global horizontal irradiance (GHI) and undertake the task of narrowing down these options for a location in Qinghai, China. We show that optimizing the configuration of the WRF model based on the type of day (sunny, partly cloudy, or cloudy) is 13.6% better than using a single best configuration for all types of days. We also show that this performance improvement holds true for a longer 3-month test period (17.8% improvement). [ABSTRACT FROM AUTHOR]

Six adaptive, short-term post-processing methods for correcting systematic errors in numerical weather prediction (NWP) forecasts of near-surface air temperatures using local meteorological observations are assessed and compared. The methods tested are based on the simple moving average and the more advanced Kalman filter. Forecasts from the rather coarse-resolution global NWP model Global Forecast System (GFS) and the regional highresolution NWP model HARMONIE are post-processed, and the results are evaluated for 100 private weather stations in Denmark. The performance of the post-processing methods differs depending on the NWP model. Overall, the combined moving average and a so-called lead time Kalman filter performs the best. The moving average was shown to be superior to a diurnal bias correction Kalman filter at removing the longer-term systematic errors for HARMONIE forecast data and comparable for GFS forecast data. Subsequent application of the lead time Kalman filter corrects for the short-term errors using the real-time forecast error. The post-processing method is adaptive and there is no need for a long record of observations or a historical archive of forecasts. [ABSTRACT FROM AUTHOR]

The societal benefits of numerical weather prediction (NWP) forecasts are most evident in populated areas. An urban representation within NWP models should provide improved forecast accuracy. Here, we present the preliminary implementation of an urban scheme within the Integrated Forecasting System (IFS) using a simplified single‐layer urban canopy model. The scheme makes assumptions of canyon geometry and considers fluxes from roads, walls, and roofs. Temperature observations were used to optimize single‐column model (SCM) parameters using the Gauss‐Newton method. Observation comparisons over six European cities, show a 2‐m temperature root‐mean‐squared error reduction from 1.85 to 1.75 K with the urban scheme. Optimized parameters were used globally at kilometric scale in a land surface model. A sensitivity experiment assuming a 100% urban world showed spatially averaged northern hemisphere 2‐m temperatures increased by 0.54 K (January) and 0.42 K (July) at night caused by changes in the albedo, emissivity, roughness, and thermal and hydrological properties. Global ∼1‐km resolution simulations using ancillary urban mapping information produce an urban heat island effect over major and minor conurbations. Only major conurbations were well represented at ∼9‐km resolution. Results from SCM simulations show a heightening of the planetary boundary layer over city sites, with the largest enhancements occurring at night in July (84 ± 48 m) caused by an increased sensible heat flux. These initial developments show the importance of a high‐resolution urban representation within NWP models. Improved parameterization and mapping will enable an online representation of energy, water, and trace gas fluxes over residential areas. Plain Language Summary: Urban areas make up only a small fraction of the Earth's surface; however, they are home to over 50% of the world's population. In these areas a phenomenon known as the urban heat island effect causes increased temperatures due to human activities, an effect often missing in weather forecasts. Forecasts, generated using computer models, consider not only the atmosphere but also the role of the land surface on the weather above. Typically these models do not include an urban map, so they miss key urban processes. We introduced a representation of urban areas to the model of the European Center for Medium‐Range Weather Forecasts. We considered several ways in which the urban environment interacts with the weather, including through changes in heat storage and treatment of rainfall. We find these developments result in a more accurate model forecast over six European cities. The model accurately predicts the increased heating observed over cities at night and some of the observed changes in the atmosphere. Future work should continue to improve the urban representation in weather and air quality/greenhouse gas models by implementing an urban scheme in operational forecasts. Key Points: An urban scheme has been introduced and optimized within the ECMWF IFS single‐column and surface‐only modelAssuming an urban world, average nighttime 2‐m temperatures increased for January (0.54 K) and July (0.42 K) in a surface only simulationUsing realistic urban cover for eight cities, PBL height in July increases by an average of 66 and 84 m for the day and night, respectively [ABSTRACT FROM AUTHOR]

Prediction of heavy/extreme rains is still a challenge, even for the most advanced state-of-the-art high-resolution Numerical Weather Prediction (NWP) modelling systems. Hydrological models use the rainfall forecasts from the NWP models as input. This study evaluates the performance of the UK Met Office Unified Model (UM) in predicting the rainfall exceeding 80th and 90th percentiles. Such high rainfall amounts occur over the Western Ghats (WGs) and North East (NE) India mainly due to the forced ascent of air parcels. Apart from the significant upgrades in the UM's dynamical core, the model features an increased horizontal grid (40–10 km) and vertical resolution (50–70 levels). The prediction skill of heavy rainfall events improves with an increased horizontal resolution of the model. The probability of detection (POD), false alarm ratio (FAR), and critical success index (CSI) are the verification metrics used. As per these metrics, model rainfall forecasts have improved during 2007–2018 (increase in CSI from 0.29 to 0.38, POD from 0.45 to 0.55, and decrease in FAR from 0.55 to 0.45). Additionally, to verify extreme and rare events, the symmetric extremal dependence index (SEDI) is also used. SEDI also shows an increase from 0.47 to 0.62 and 0.16 to 0.41 over WGs and NE India during the study period, suggesting an improved skill of predicting heavy rains over the mountains. The improved forecast performance is consistent and relatively higher over WGs than over NE states. [ABSTRACT FROM AUTHOR]

To circumvent the scalability problem due to global communication involved in spectral transforms, a vertical‐slice version of the dynamical core where all calculations are performed in grid‐point space has been built for AROME, Météo‐France's operational limited‐area model. It is shown in an idealized but nevertheless physically relevant framework that, despite this major change, it is possible to keep the other main characteristics of the model (constant‐coefficient semi‐implicit scheme, semi‐Lagrangian transport scheme, A‐grid, mass‐based coordinate, etc.). A Krylov solver is used to solve the implicit problem. Using the solution given by the spectral model as a reference in terms of quality and required accuracy in an operational context, the chosen parameters of the Krylov solver are carefully tuned to maximize its convergence speed. The Krylov solver consists of several applications of a sparse operator, whose stencil is similar to that of the operator applied during the small time step of split‐explicit schemes traditionally used in HEVI (Horizontally Explicit/Vertically Implicit) models. HEVI models are considered particularly scalable in the current parallelization paradigm and are used as a reference for scalability in this study. Results show that a fast convergence can be reached, such that only a few applications of the sparse operator are required. This suggests an improvement in scalability compared to the spectral version. Experiments have been formulated regardless of the future strategy considered for temporal and spatial resolutions, i.e., increasing spatial resolution while either maintaining a high Courant number or reducing it. [ABSTRACT FROM AUTHOR]

A new method has been developed to directly assimilate volume scans of radar reflectivity data with 4D‐Var in the Met Office Unified Model. The method has been demonstrated in the convective‐scale hourly‐cycling UKV forecast model. Reflectivity observations from 18 C‐band radars in the British Isles are assimilated. This article describes the method of observation processing and quality control, the observation operator, and assimilation method. The assimilation method uses a minimum threshold rainwater content in the forward operator to give sensitivity to reflectivity observations where there is no rain in the background. Furthermore, the use of the Huber norm in the observation penalty function allows the use of observations with large innovations in the assimilation. A change was made to the precipitation efficiency in the microphysics scheme of the linear perturbation forecast model to ensure stability of the scheme. A case‐study is presented which demonstrates how the inclusion of reflectivity observations enhances convergence through analysis increments to the wind field, leading to improvements to the location of convective precipitation features in the forecast. Two‐month trials for summer and winter seasons demonstrate significant improvements to rain forecasts in the nowcasting range. [ABSTRACT FROM AUTHOR]

The numerical weather prediction (NWP) of fog remains a challenge, with accurate forecasts relying on the representation of many interacting physical processes. The recent Local And Non-local Fog EXperiment (LANFEX) has generated a detailed observational dataset, creating a unique opportunity to assess the NWP of fog events. We evaluate the performance of operational and research configurations of the Met Office Unified Model (MetUM) with three horizontal grid lengths, 1.5 km and 333 and 100 m, in simulating four LANFEX case studies. In general, the subkilometre (sub-km) scale versions of MetUM are in better agreement with the observations; however, there are a number of systematic model deficiencies. MetUM produces valleys that are too warm and hills that are too cold, leading to valleys that do not have enough fog and hills that have too much. A large sensitivity to soil temperature was identified from a set of parametrisation sensitivity experiments. In all the case studies, the model erroneously transfers heat too readily through the soil to the surface, preventing fog formation. Sensitivity tests show that the specification of the soil thermal conductivity parametrisation can lead to up to a 5-hr change in fog onset time. Overall, the sub-km models demonstrate promise, but they have a high sensitivity to surface properties. [ABSTRACT FROM AUTHOR]