ABSTRACT Afghanistan is highly vulnerable to hydrological‐related hazards. Following years of civil unrest and lack of investments, the gauge network in Afghanistan has declined severely, and operational applications mainly depend on publicly available gridded precipitation datasets. Despite this pressing need, little has been published on the quality of widely used gridded precipitation datasets for the region. This study assessed CHIRPS V2.0, IMERG V07B, EAR5, MSWEP V2.8, and GMCP against a spatially interpolated rain gauge dataset for the period 2008–2020. The main findings include: (1) all the five precipitation datasets demonstrate generally similar climatological spatial patterns with increasing precipitation volumes from the southwest to the northeast; (2) GMCP demonstrated the overall highest skills in terms of total annual amounts, error statistics, and the seasonal variations, followed by CHIRPS, MSWEP, IMERG, and ERA5; (3) The largest biases are evident over the northeastern regions, where the ground reference data are in low quality and sparce. The driven datasets and/or algorithm schemes are vital for the performances of the five gridded datasets, that is, GMCP is robust in removing false‐alarmed precipitation pixels by introducing morphological theory to optimally identify the precipitation occurrence merging weights, both climatological estimates and near‐real‐time calibrations using plenty ground observations are comparably important for CHIRPS, the sparse local observations greatly limit the abilities to obtain the optimal multisource merging weights for IMERG‐Final and MSWEP V2, and the interactions between the dynamical models and the evolving observing system still exhibit nonzero and often substantial random errors in generating ERA5. Results of this study provide promising references for selecting the optimal datasets in precipitation‐related research and applications.
ABSTRACT An adaptive algorithm has been developed using ground‐based high‐resolution aerosol backscatter profiles from the Vaisala CL61 ceilometer to retrieve planetary boundary layer height (PBLH). Datasets spanning more than 3 years (2022–2025) from the U.S. National Capital Region (NCR) were comprehensively analyzed. We focused on examining the PBL dynamics representing urban and suburban environments in the National Capital Region. Our findings confirm that the daytime PBLH in both environments follows a typical solar‐driven seasonal cycle. However, the nocturnal boundary layer (NBL) from both sites showed more complex behavior. Although the urban heat island (UHI) related storage may influence nighttime stability, the observed urban‐suburban contrast could not be explained by the UHI effects alone. The suburban site frequently showed elevated nighttime PBLH, suggesting an additional role of mechanical turbulence and land‐atmosphere coupling. To understand the underlying mechanism of this behavior, we analyzed PBL development, particularly after precipitation. Results suggest that following precipitation, PBLH was elevated in the suburban environment. Mean suburban nighttime PBLH was lower 4–5 days after rainfall than in the earlier post‐rainfall intervals, although the change was not monotonic. These patterns suggest a possible influence of land‐surface conditions, but the contributions of soil moisture and evapotranspiration remain uncertain. It appears that soil moisture capacity relative to precipitation may play an important role in the evolution of suburban PBLH. Furthermore, the High‐Resolution Rapid Refresh (HRRR) model was used to compare with observational PBLH and evaluate any systematic biases in the model. Results suggested that the model tends to overestimate PBLHs under convective conditions and significantly underestimate them under stable nocturnal conditions. ERA5 reanalysis latent heat flux data from 2022 to 2025 were further analyzed to investigate the potential role of surface energy exchange in explaining the consistently higher PBLHs observed at the suburban site. The results indicate that differences in energy exchange between urban and suburban environments may contribute to the observed site‐dependent variability in PBLH. Results from this study also demonstrate that the newly developed adaptive algorithm with the CL61 ceilometer may be used for long‐term PBL monitoring and to further fine‐tune PBL parameterizations in high‐resolution meteorological models.
ABSTRACT We combined high‐resolution (1–4 km) numerical simulations with nonlinear balance diagnostics, cross‐Lagrangian Rossby number analysis, shear‐instability metrics, vertical cross sections, and soundings to identify the multiscale dynamical processes governing the genesis of a largeamplitude solitary Mesoscale Gravity Wave of Depression (MGW D ). Results indicate that MGW D initiation arises from a multiscale sequence involving the presence of terrain‐modulated pressure, geopotential thickness anomalies, upper‐level jet imbalance, and localized shear instability. Terrain sensitivity experiments reveal a critical role of terrain in modulating multiscale imbalances during split frontogenesis. Based on the analysis presented in the study, a four‐step mechanism is proposed, where (1) terrain‐induced splitting of the synoptic cold front over the Sierra Madre Oriental Mountain range located in northeastern Mexico enhances leeside cold‐air advection and mesoscale cold‐air damming, leading to the development of a downstream warm ridge; (2) the resulting difference in the southwest–northeast‐oriented atmospheric column thickness, enhances the geopotential height field gradient relative to the wind structure, intensifying and contracting the upper‐level trough. The remaining imbalance on timescales short relative to the inertial period consequently results in rapid balance adjustment, enhanced divergence tendencies, and the emergence of an unbalanced jetlet in the mid‐troposphere; (3) descent associated with the jetlet circulation results in subsidence warming and promotes mesoscale mass flux divergence at the surface. The downward motion upon impulsively hitting the stable cold‐air layer accompanying the cold sector of the extratropical cyclone, results in the observed MGW D and associated surface pressure falls; (4) furthermore, the mid‐ to upper‐level frontal splitting due to the terrain reinforces the solitary wave signal and amplifies the inverted surface trough. These results identify a previously underexplored pathway linking terrain‐modulated frontal evolution and jet imbalance to MGW genesis, with implications for improving operational forecasting and numerical weather prediction, which requires assessing whether such events consistently follow the proposed mechanism.
ABSTRACT Model development cycles produce large volumes of verification output, often summarised through scorecards based on univariate measures such as root‐mean‐square error (RMSE). Although useful as broad indicators of improvement or degradation, scorecards can be difficult to interpret across many variables, lead times and successive trials. This study examines whether dissimilarity‐based methods can provide a more compact summary of performance changes through a development cycle. Two complementary methods are applied to transformed RMSE scorecard components from 26 trials: five controls and 21 test configurations. First, multidimensional scaling (MDS) is used to produce two‐dimensional representations of relative similarity between trials at each lead time. Second, distribution‐based two‐sample statistics, including Kolmogorov–Smirnov, Cramer‐von Mises, Anderson‐Darling, and Wasserstein‐based measures, are applied to compare whole‐trial distributions of transformed RMSE. A z ‐score transformation places variables with different physical units on a common scale. The methods are demonstrated using a development cycle that introduces a new dynamical core on an unstructured cubed‐sphere mesh while retaining the same science configuration as the previous model. The cycle had the unusual aim of achieving broadly comparable performance rather than clear improvement. MDS places the most refined atmosphere‐only data‐assimilation trials close to their controls and reveals structure associated with season and data assimilation. The distribution‐based comparisons likewise show little evidence of dissimilarity between the final trials and relevant controls while indicating where the largest differences occur. These results suggest that dissimilarity‐based summaries can complement traditional scorecards by exposing relationships, convergence and divergence in a form that is easier to inspect and interpret.
ABSTRACT Accurate soil moisture estimates are essential for effective management and operational planning of flood and drought events, both of which are intensifying under a changing climate. However, soil moisture values derived from land surface models often exhibit significant deviations from in situ observations. Data assimilation combines model information with observations to enhance prediction accuracy. Previous studies have typically focused on either estimating the initial soil moisture state or optimising Pedotransfer Function (PTF) constants (or parameters), which link soil texture to the hydraulic properties of land surface models. In contrast, this study introduces a novel extension by performing joint state‐parameter assimilation for the JULES model, simultaneously optimising both PTF constants and initial soil moisture states within an augmented framework. Using an Ensemble–Variational hybrid data assimilation scheme, we ingested field‐scale soil moisture observations from the Cosmic‐ray Soil Moisture Monitoring Network across 16 diverse sites in the UK. The joint state‐parameter assimilation improved average Kling–Gupta Efficiency by approximately 113% relative to the open‐loop, 105% relative to state‐only and 9.6% relative to parameter‐only experiments. These results highlight the strong potential of the proposed framework for enhancing land‐surface reanalysis, numerical weather prediction and hydrometeorological forecasting of floods and droughts.
