Meteorological Applications最新文献

This study employed high-resolution (1 × 1 km) multisource fusion data (HCLDAS) and observational data from 190 automatic weather stations to analyze summer temperature variations across 12 altitude levels in the Hotan area from June to August 2023. Statistical methods, including root mean square error (RMSE) and temperature accuracy rates (TT1, TT2), were applied to validate data reliability and investigate spatiotemporal patterns. Key findings include: (1) Data Validation: HCLDAS demonstrated high accuracy, with a mean RMSE of 0.42°C and temperature accuracies of 98.15% (≤ 1°C) and 99.08% (≤ 2°C), confirming its suitability for complex terrains. (2) Altitude-Dependent Trends: High elevations (≥ 4500 m): Continuous warming from July to August (+0.37°C to +0.96°C), driven by glacier-albedo feedback (e.g., Muztagh Ata retreat) and weakened westerlies enhancing thermal forcing, elevating the 0°C isotherm. Mid-elevations (2000–4500 m): Sharp vertical cooling (−18.21°C total) but significant June–July warming (+1.24°C to +2.96°C). Low elevations: July–August cooling (−0.07°C to −1.05°C) due to cold air drainage and oasis effects (evaporation/dust reflection). (3) Diurnal Variability: Maximum daily temperature range (12.6°C) occurred at 1300–1500 m (arid landscapes), while the minimum (6.08°C) was observed at 4000–4500 m (rocky terrain). (4) Threshold Analysis: ≤ 0°C grids (38.51% of total) concentrated above 2500 m, while ≥ 35°C grids (55.59%) dominated below 3000 m, with cumulative hours increasing at lower altitudes. The results provide a scientific basis for high-temperature monitoring, snowmelt flood warnings, and optimized meteorological infrastructure in arid, high-altitude regions.

The Little Dry Season (LDS), a distinct feature of the West African Monsoon system, separates the major and minor rainfall seasons in the Guinea Coast's bimodal rainfall regime. Despite its significant socio-economic implications, the LDS is poorly understood in terms of its historical patterns, key drivers, and future projections. In this study, we analyze the historical and future patterns, variability and drivers of the LDS, pinpointing August as the month when its characteristics are most prominent. The historical period comprised data from 1990 to 2020, and the projection data was split into three climate regimes namely near-future (2011–2040), mid-future (2041–2070) and far-future (2071–2100). We identify Sea Surface Temperature (SST) and Top of the Atmosphere Outgoing Longwave Radiation as critical surface drivers for detecting and characterizing the LDS. Subsequently, we validate CMIP6 climate models against CHIRPS observational data, applying bias correction to enhance their accuracy in simulating LDS rainfall. Future LDS patterns are projected under three Shared Socio-economic Pathways (SSPs), revealing increased light rainfall events, significant spatial variability, and strong scenario dependence, particularly under SSP5-8.5. These findings underscore the need for integrated climate adaptation strategies and highlight the critical importance of global mitigation efforts in shaping future climate risks in this sensitive region. Understanding and preparing for shifts in LDS patterns is crucial for sustainable development and resilience in West Africa.

Summer precipitation over the northern part of the Korean Peninsula (SP-NPKP) is critical for water resources, agriculture, and disaster prevention. This study aims to detect suitable atmospheric circulation indices for annual prediction of SP-NPKP and to evaluate their predictive skill. We used 77 years of data from 1948 to 2024, including NCEP/NCAR reanalysis variables and observed summer precipitation from 37 stations. The study is based on the finding that 1-year lag correlations between selected indices and SP-NPKP generally exceed concurrent correlations. We analyzed linear trends of SP-NPKP, sea level pressure over Asia, 700-hPa vorticity anomalies, and Arctic Oscillation indices. Using the ‘area shift’ experiment, we identified optimal domains for sea level pressure anomalies over Asia and the North Pacific, yielding effective predictors: the SLP anomaly index over central Eurasia (SLPAI-1), that over the Okhotsk Sea (SLPAI-2), the 700-hPa vorticity anomaly index (VORAI-700), the 500-hPa temperature anomaly index (TAI-500), and the leading SLP principal component (SLP-PC1). Annual predictions were performed using principal component regression (PCR) and backpropagation neural network (BPNN) models. Based on 5-fold cross-validation, PCR showed limited skill with R² = 0.1628, RMSE = 151.57 mm, and MAE = 124.22 mm, while BPNN demonstrated significantly superior performance with R² = 0.4031, RMSE = 119.81 mm, and MAE = 103.15 mm. This confirms that neural networks better capture the nonlinear dynamics of regional precipitation. Our study provides a novel, data-driven framework for identifying region-specific predictors, offering valuable insights for improving operational seasonal prediction systems in East Asia.

This study addresses the critical issue of missing marine wave observation data, including significant wave height, mean wave period, and mean wave direction, which are essential for oceanographic analyses and marine operations. An imputation model based on the Weighted K-Nearest Neighbors (WKNN) algorithm is proposed, using the square of wave height as the primary input feature. This height-squared formulation, physically motivated by wave energy density being proportional to the square of wave height, has been shown to improve imputation accuracy for missing wave data, particularly when combined with standardization preprocessing. It outperforms the more common but less effective practice of using unsquared wave height values. The model is evaluated using real-world data from four buoys deployed in the northeastern waters of Taiwan. This improvement raises overall data completeness from 63.1% to 98.9%. The model yields physically plausible estimates, demonstrating strong performance in smooth to moderate WMO sea states. In rough-and-above regimes, however, the imputation results can be slightly conservative, including during typhoons. Notably, the proposed approach remains effective even when data from up to half of the buoy stations are unavailable. By generating high-quality imputed data, the model directly enhances the reliability of real-time marine monitoring and provides robust support for wave climate analysis and marine energy assessments. The results highlight the computational efficiency, robustness, and practical applicability of the WKNN algorithm in operational oceanographic systems.

Extreme weather far beyond what has been experienced in recent memory can be especially dangerous and costly. Proactively identifying locations at high risk of experiencing unprecedented weather can assist with disaster preparedness. Such locations can be referred to as having soft records, meaning the most extreme event in observational records is not particularly severe compared to what is possible. In previous studies, the systematic identification of soft records over a large spatial domain involves applying extreme value analysis to gridded observational or reanalysis data. A limitation of these studies is the small sample size, which we propose can be addressed by adapting the UNprecedented Simulated Extremes using ENsembles (UNSEEN) approach that is commonly used to estimate event likelihood in the aftermath of isolated unprecedented events. The UNSEEN approach makes use of seasonal or decadal forecast/hindcast ensembles, which provide a large sample of plausible events over recent decades. To demonstrate the utility of applying the UNSEEN approach to a large spatial grid, we assessed record daily maximum temperatures across Australia using gridded observations and data from 10 different decadal forecasting systems. The observation-based results highlighted broad areas of soft records in the south-east of mainland Australia, extending north into south-west and western Queensland. The UNSEEN-based analysis also identified soft records in western Queensland, but not in the south-east where the underlying positive trends in extreme temperature were far less severe in the models than in observations. We suggest that the use of large model ensembles (i.e., an UNSEEN-based approach) can complement an observation-based approach to identifying soft records over large gridded spatial domains.

We apply a tropical cyclone (TC) track classification method to the subseasonal TC predictions issued by the European ensemble prediction model (EPS) on the Southwest Indian Ocean. This track typology is based on a subdivision of the basin into three areas. Each simulated storm track is assigned a three-digit code, where each digit represents the area of genesis, the westernmost position, and the easternmost position, respectively. To account for the intensity bias in the simulated TCs, we conduct sensitivity tests that result in lowering the tropical storm wind threshold to 29kt and filtering out systems with a lifetime maximum intensity lower than 34kt. The model skill is evaluated against the performance of a 1-month moving climatology, to validate its ability to capture intra-seasonal variations in TC activity and favoured track typology. Results show an overestimation of occurrence probabilities for all track types, with the central part of the basin and the Mozambique Channel being the regions most affected. These limitations confine the raw model's skill to the second week (W2, i.e., +7 to +14 days) of forecast only. However, when evaluating the model on its capacity to assign a track type to TC genesis known to be valid at the basin scale, the EPS exhibits a 20% performance gain for W2 and a 5% gain at W3 and W4, compared to the moving climatology. These results demonstrate that when TC forecasters consider an EPS genesis prediction to be reliable, they can leverage the corresponding track type predictions to characterise TC risk more precisely, even a month in advance. Furthermore, aggregating track types based on their likelihood of impacting inhabited areas within the basin further enhances predictive skill. An impact-based forecasting product is derived from this work and will be evaluated in operations by the TC forecasters in La Reunion.

This study investigates the accuracy of GLDAS-Noah-2.1 daily minimum soil temperature (DMST) at 0–10 cm (shallow) and 40–100 cm (deeper) depths in Khorasan Razavi, Iran, using 5 years (2008–2012) of data from 13 synoptic stations. Initial evaluations using statistical metrics revealed significant discrepancies, with GLDAS tending to overestimate DMST. The initial raw GLDAS data showed a considerable systematic error, with a bias ranging from 1.96°C to 6.0°C in the shallow profile and from 0.58°C to 6.62°C in the deeper soil profile. On average, the shallow layer performed poorly, yielding an RMSE of 4.82°C and an average bias of 3.85°C, while the deeper layer showed an average RMSE of 2.72°C and a bias of 2.14°C. To mitigate these biases, a K-nearest neighbors (KNN) supervised machine learning algorithm was employed and optimized through grid search. The KNN model dramatically enhanced performance for both layers. For the shallow depth, the average RMSE was reduced to 2.56°C (a ~47% reduction), and the average bias was reduced to 0.06°C. For the deeper layer, the average RMSE was reduced to 1.30°C (a ~52% reduction), and the average bias was reduced to 0.00°C. Furthermore, the Nash–Sutcliffe efficiency (NSE) improved from an initial average of 0.75 and 0.86–0.92 and 0.97 for the shallow and deeper layers, respectively. Post-correction, the model achieved a “Very Good” performance rating for all stations, with the average percent bias (Pbias) falling to 0.37% (shallow) and −0.08% (deeper). The results underscore the efficacy of machine learning-based bias correction in enhancing the reliability of GLDAS datasets for regional climate and agricultural applications.

Tropical cyclone (TC) Doksuri (2023) exhibited a sudden northward turn over the northeastern part of the South China Sea (SCS). However, most global models failed to capture such track change. The US National Centers for Environmental Prediction Final Analysis (FNL) data and The International Grand Global Ensemble (TIGGE) data were therefore used to study the underlying mechanisms for the sudden track change and the factors leading to the track forecast errors of different global models so as to give some insight for the forecasters in predicting such TC track change and global model developers in modifying the model physics. The non-linear advection of the vorticity of the asymmetric winds associated with the monsoon trough over the SCS and that of the symmetric wind of the TC resulted in the sudden northward turn of the TC track. However, the strength and the eastward extension of the monsoon trough were underpredicted, leading to a westward-moving track without a sharp northward turn. On the contrary, if the strength of the monsoon trough was overpredicted, the environmental steering was over-altered, resulting in an early northward turn. The intensity and outer wind structure of the TC also played important roles in the monsoonal interaction and thus the track forecast errors.

Dry intrusions are coherent airstreams that originate from the upper troposphere, or lower stratosphere, and descend towards the surface where they can influence the dynamics of mid-latitude weather. Notably, the occurrence of these airflows with atmospheric fronts and extratropical cyclones can exacerbate their impacts, leading to increased precipitation and stronger surface winds. Therefore, it is of interest to understand how dry intrusions may respond to our changing climate. Traditional identification methods with Lagrangian trajectory analysis, however, cannot always be applied to climate projection data due to the computational cost of this approach and a lack of necessary available data from climate models. This research explores the alternative application of image segmentation concepts to build a machine learning classification model for dry intrusion outflows in the North Atlantic. A U-Net convolutional neural network (CNN) is trained to predict the presence of dry intrusion objects from ERA5 atmospheric data, including temperature and relative humidity. With a catalogue of labelled dry intrusion objects calculated from trajectory analysis as predictands, the ability of this CNN to identify individual dry intrusion footprints, capture their size and shape, and recreate long-term climatologies is evaluated with Matthew's correlation coefficient and intersection over union. Compared with a multiple logistic regression model, the CNN outperforms across all metrics and compares more favourably with the target data. However, the CNN struggles to predict small dry intrusion signatures and limitations are encountered outside of the spatial training domain in a high-impact case study. Despite this, these results provide proof-of-concept for an alternative way to identify dry intrusion outflows that uses less data, is fast and easy to implement, and could be utilised to study the possible futures of dry intrusions and extreme mid-latitude weather.

Near-surface observations can suffer from significant representativeness errors, especially for Numerical Weather Prediction (NWP) at lower resolution in global applications. Therefore, in Data Assimilation (DA), many operational centers have long been reluctant to assimilate them (e.g., the European Center for Medium-range Weather Forecast, ECMWF, started assimilating all 6-h screen-level temperature reports only in 2024). For forecast verification, some studies advocate that we should not rely on them and use only verification against our own near-surface analyses. At Environment and Climate Change Canada (ECCC), both temperature and humidity observations from SYNOPs have been assimilated in our global NWP system for more than two decades and, in June 2024, METARs have been added following some positive impacts found only when comparing forecasts against near-surface observations. To shed light on the impact of the assimilation of screen-level observations, in this study we present an evaluation of the impact of removing the assimilation of all screen-level temperature and humidity observations using various verification references: the NWP forecasts were evaluated against radiosondes and surface observations, independent (ECMWF) analysis, our own analysis and surface analysis. Results show that, despite the lack of a proper estimation of representativeness errors in the DA approach, the assimilation of screen-level temperature and humidity leads to forecast improvements that can be detected from the verification against independent measurement sources, here radiosondes and ECMWF upper-air analyses. Verification against own analyses, for both upper-air and screen-level variables, led instead to opposite and misleading conclusions. In fact, the removal of assimilated screen-level temperature and humidity measurements renders the NWP forecast more similar to the own analysis, therefore leading to better scores but detachment from the observed world.