High-resolution Numerical Model Research Articles

Instabilities in the Bottom Boundary Layer Reduce Boundary Layer Arrest and Stir Boundary Layer Water Into the Stratified Interior

AbstractAn along‐isobath current in stratified waters leads to a bottom boundary layer. In models with no alongshore variation, cross‐isobath density transport in this bottom boundary layer reduce the velocity in the bottom boundary layer via thermal wind, and thus the bottom friction experienced by the current above the boundary layer—this is bottom‐boundary‐layer arrest. If, however, alongshore variation of the flow is allowed, the bottom boundary layer is baroclinically unstable. We show with high resolution numerical models that these instabilities reduce this arrest and allow bottom friction to decelerate the flow above the bottom boundary layer when the flow is in the Kelvin wave direction (so that the bottom Ekman transport is downwelling). Both the arrest of the bottom boundary layer and the release from this arrest are asymmetric; the friction experienced by flows in the direction of Kelvin‐wave propagation (downwave) is much greater than flows in the opposite direction. The strength of the near bottom currents, and thus the magnitude of bottom friction, is found to be governed by the destruction of potential vorticity near the bottom balanced by the offshore along‐isopycnal transport of this anomalous potential vorticity. A simple model of this process is created and used to quantify the magnitude of this effect and the resulting reduction of arrest of the bottom boundary layer.

Modelling the Oceanic Advection of Pollutants Spilt Along with the Northwest Passage

ABSTRACT The Arctic sea ice is dramatically retreating in concentration, thickness, and duration. The larger and longer-lasting periods of open water will likely lead to increase trans-Arctic ship traffic, which then increases the risk of accidents and of pollutant spills. In this study, we focus on the potential oceanic circulation pathways of pollutants that may be spilt along with the Northwest Passage in the Canadian Arctic. We used a high-resolution numerical model and a Lagrangian particle tracking tool to simulate the advection of pollutants released in and within proximity to the Canadian Arctic Archipelago. We released 5000 virtual particles over 24 main release sites every 10 days during the operating season (June–October) for 12 years (2004–2015). For each simulation, we examined the circulation pathway and computed particles’ spreading area, distances travelled, subsurface spread, and the variability and uncertainty of their distribution during the two-year simulation duration. We analysed these factors with respect to the dominant oceanic circulation of where the particles were initially seeded and the role of atmospheric circulation and were able to identify three main circulation regimes and eight small-scale regimes. This study highlights the role of oceanic advection in the spreading of particles and determines that particles released in the eastern study area exhibited the largest spreading area as the majority propagated into the North Atlantic Ocean rapidly.

Transport Structure of the South Atlantic Ocean Derived From a High-Resolution Numerical Model and Observations

The South Atlantic Ocean plays an important role in the Atlantic meridional overturning circulation (AMOC), connecting it to the Indian and Pacific Oceans as part of the global overturning circulation system. Yet, there are still open questions regarding the relative importance of the warm water versus cold water sources in the upper limb of the AMOC and on the detailed circulation pathways of the North Atlantic Deep Water (NADW) in the lower limb. These questions are addressed using model outputs from a 60-year, eddying global ocean-sea ice simulation that are validated against observations. We find that the Pacific Antarctic Intermediate Water (AAIW) plays a role in setting the temperature and salinity properties of the water in the subtropical South Atlantic, but that the upper limb of the AMOC originates primarily from the warm Indian water through the Agulhas leakage (9.8 Sv of surface water + 3.5 of AAIW) and that only a relatively small contribution of 1.5 Sv colder, fresher AAIW originates from the Pacific Ocean. In the lower limb, the NADW flows southward as a deep western boundary current all the way to 45°S and then turns eastward to flow across the Mid-Atlantic Ridge near 42°S before leaving the Atlantic Ocean, although there is clockwise recirculation in the Brazil, Angola, and Cape Basins.

Hydrokinetic Power Resource Assessment in a Combined Estuarine and River Region

The worldwide river and tidal hydrokinetic power potential is considerable. Harnessing such potential could allow the generation of a significant amount of sustainable electricity for local uses. To the present, most studies on hydrokinetic power have focused on large-scale commercial technology development, large tidal farms planning, and high-intensity resources assessment. Reduced attention was oriented towards investigating possibilities for small to medium-size hydrokinetic plants. However, given the characteristics of rivers and estuaries, in most cases, relevant hydrokinetic power exploitation possibilities exist regardless of the dimensions of the region considered. The planning of small to medium-size hydrokinetic plants for various aspects differs from larger developments. In the present work, a method for assessing the hydrokinetic resource is proposed and applied to the case study of the Douro waterway, which is characterized by moderate flow speeds and limited water depths. A high-resolution shallow-water numerical model is set up using ocean and river inflow boundary conditions. The flow velocities are estimated for the neap-spring period for different freshwater discharges. The spots presenting the highest annual hydrokinetic power average were identified, maximum flow speeds of about 1 m/s were found, and an annual mean power of 0.4 kW/m2 was estimated, indicating that prospects for hydrokinetic energy harvesting exist.

High-efficiency and high-resolution numerical modeling for two-dimensional infiltration processes, accelerated by a graphics processing unit

High-efficiency and high-resolution numerical modeling for two-dimensional infiltration processes, accelerated by a graphics processing unit

Marine Online Platforms of Services to Public End-Users—The Innovation of the ODYSSEA Project

Recently, various Earth Observation Networks (EONs) have been designed, developed and launched by in-situ, on-site and off-site collected data from fixed and moving marine sensors and remote sensing (RS) satellite data. This information can significantly help a wide range of public and private end-users better understand the medium- and high-resolution numerical models for regional, national and global coverage. In this context, such EON core services’ operational numerical data can be seen of the growing demand result for marine sustainability development of developing countries and the European Union (EU). In this case, marine platforms can offer a wide range of benefits to users of human communities in the same environment using meticulous analyses. Furthermore, marine platforms can contribute to a deeper discourse on the ocean, given the required regulations, technical and legal considerations and users to a common typology using clear scientific terminology. In this regard, firstly, the following six steps have been used to develop a better understanding of the essential data structure that is commensurate with the efficiency of the marine end-user’s service: (1) steps and challenges of collecting data, (2) stakeholder engagement to identify, detect and assess the specific needs of end-users, (3) design, develop and launch the products offered to meet the specific needs of users, (4) achieve sustainable development in the continuous provision of these products to end-users, (5) identify future needs and challenges, and (6) online platform architecture style related to providing these products to end-users. Secondly, the innovation of the ODYSSEA (Operating a Network of Integrated Observatory Systems in the Mediterranean Sea) platform project has been evaluated and reviewed as a successful project on marine online platforms to better understand how marine online platforms are being used, designed, developed and launched. The ODYSSEA platform provides a system that bridges the gap between operational oceanographic capabilities and the need for information on marine conditions, including for the end-user community. The project aims to develop a fully integrated and cost-effective cross-platform, multi-platform network of observation and forecasting systems across the Mediterranean Sea.

Sea-level extremes of meteorological origin in the Red Sea

Severe weather events and the resulting sea-level extremes are considered among the greatest threats to coastal environments. Understanding this coastal hazard is vital for various coastal developments and planning activities. This study examines the sea-level extremes of meteorological origin in the Red Sea. A high-resolution (approximately 500 m) depth-averaged numerical ocean model forced by 5 km dynamically downscaled meteorological fields was implemented to hindcast the sea level of the Red Sea over 37-year, 1980–2016. The model was first validated with observations, showing good agreement with available data. The hindcast model outputs were then analyzed to describe the spatiotemporal features of the sea-level extremes in the Red Sea. The sea-level extremes in the Red Sea developed in response to wind variability over the southern Red Sea and exhibited a basin-wide impact. The magnitudes of the maximum sea levels reached approximately 0.30–0.50 m inside the Red Sea basin, with the highest values (0.85 m) in the Gulf of Suez. The seasonal distribution of the extremes suggests that these are frequent during the winter months (January–March). Assessment of long-term changes in the annual maxima, 99th and 95th percentiles of hourly sea levels indicate no significant trends over the study period.

A high-resolution numerical well-test model for pressure transient analysis of multistage fractured horizontal wells in naturally fractured reservoirs

In this study, a high-resolution numerical well-test model based on hybrid discrete fracture methods for pressure transient analysis of multistage fractured horizontal wells in naturally fractured reservoirs is proposed. Given the complexity of gridding for a numerical transient model, a new type of hybrid grid system honoring complex geometries of high-velocity transient flow near wellbore and hydraulic fractures is developed. Meanwhile, the methods of identifying fracture intersections in 3D unstructured space, transmissibility corrections and stochastic fracture network generation algorithm are proposed to greatly improve the accuracy and efficiency of capturing the pressure transients in complex fracture networks. After comparison with the well-test software KAPPA, it shows that the model can not only accurately reproduce the complex 3D convergence of flow regimes but also has a much better computational performance. Then, the type curves of a multistage fractured horizontal well with complex fracture networks and the sensitivity analysis of nonuniform properties of fracture networks are developed and analyzed in detail. The results show that the early-time horizontal radial flow and dip are positively related to the properties of natural fractures, and the fluid contribution and pressure distribution mainly depend on fracture connectivity rather than fracture number. The orthogonal fracture connections can result in stronger fluid supply and less fracture interference. Besides, the fracture linear flow and first/second bilinear flow are mainly affected by hydraulic and natural fractures, respectively. Finally, it proves that the proposed model can be much more efficient than KAPPA in history match for well-test data by an actual fractured field case.

Ground- and Satellite-Based Evaluation of WRF Snowfall Prediction

This study performed 4-day numerical integration in 1-hour intervals using the Weather Research and Forecasting (WRF) model for four major cases of heavy snowfall that occurred from 2020 to 2021. The model-predicted snow depth data were compared with the ground-observed snow depth and the satellite-observed snow cover data and then were statistically verified. The scalar verification results for ground data from the four cases showed a root–mean–square error of 2.55-16.67 cm and a correlation coefficient of 0.48-0.80, whereas the verification results with satellite data showed the correlation coefficients of 0.38-0.60. For categorical verification, using a threshold value of a snow depth exceeding 5 cm, the proportion correct was 90% or higher for ground observations of each case. In addition, in the satellite categorical verification, when the threshold value of the Snow Cover Fraction (SCF) exceeds 0.5, the proportion correct was 50% or more. These results are meaningful because the model snow depth verification methods were devised strategically for the first time using both the snow depth data of the mesoscale ground observation networks and ultra-high-resolution Sentinel-2 satellite data currently available in Korea. The findings of this study will contribute to the development of a high-resolution numerical prediction model and its verification methodology for snowfalls in the Korean Peninsula, eventually leading to increased prediction accuracy and reduced snow damage.

A High-Resolution Numerical Model of the North Aegean Sea Aimed at Climatological Studies

A new, high-resolution model for the northern part of the Aegean Sea, aimed primarily at climatological research (relaxation and data assimilation-free climate simulations), is hereby presented, along with the results of a 28-year-long simulation covering the period from 1986 to 2013. The model applied is the Regional Ocean Modelling System (ROMS). A significant improvement over previous models of the Aegean introduced in this work is the replacement of parameterizations of the Dardanelles exchange by a fully three-dimensional simulation of the flow in the Strait. The incorporation of part of the Marmara Sea in the model domain enables the interaction with other regional climate simulations, thus allowing climatic variability of the exchange of the Mediterranean and Black Seas. An extensive validation is carried out comparing the model output with all the available observations from several different platforms, i.e., satellite sea surface temperature and height, T/S profiles from R/V ships, and HF radar surface currents velocity. We focus on the model’s ability to reproduce, to some extent, the distinct thermohaline features and circulation patterns that characterize this specific area of the Mediterranean Sea. Our findings, after comparing simulation results with all the available observations, revealed the model’s sufficiency to simulate very adequately the complex hydrology of the North Aegean Sea, and the model’s ability to reproduce incidents of deep-water formation that took place in the region in previous decades during the Eastern Mediterranean Transient (EMT).

Tidal stream energy potential in the Shannon Estuary

The tidal and river in-stream energy resource in the Shannon Estuary (W Ireland) is investigated using of high-resolution numerical modelling and spatial analysis. Although freshwater discharges are large, their influence on the available resource is found to be all but negligible, the tide being the main driver of estuarine circulation. The Tidal Stream Exploitability (TSE) index is adapted to the analysis of estuaries with non-depth-limited areas (TSEndl), such as the Shannon Estuary, and then used to select the hotspots with potential for a tidal stream farm. For this purpose, a new depth penalty-limiting function is defined to avoid overestimating the available energy potential in areas with depths greater than those required for tidal energy converter operation. Seven hotspots are identified based on the revised index. The approach followed in this study illustrates the applicability of high-resolution numerical modelling and spatial analysis for identifying the most appropriate areas for tidal stream energy conversion. Finally, the potential of tidal stream energy to contribute to the much-needed decarbonisation of the energy mix in Ireland is emphasized.

Simultaneous intruding of mafic and felsic magmas into the extending continental crust caused by mantle plume underplating: 2D magmatic-thermomechanical modeling and implications for the Paleoproterozoic Karelian Craton

Available data suggest that the breakup of the Neoarchean Kenorland supercontinent at 2.5–2.4 Ga was likely triggered by a large mantle plume upwelling that caused significant magmatism. Here, we present 2D high-resolution magmatic-thermomechanical numerical models of extension of the continental crust underplated by a hot mantle plume material. Using this model, it is demonstrated that mantle plume underplating generates a large amount of mafic melt by decompression melting. This melt penetrates into the extending continental crust along normal faults thereby forming multiple generations of mafic dyke-like intrusions along normal faults. In case of extension velocity of 0.2–1 cm/yr, lower crustal heating and hot mafic melt emplacement may cause partial melting of the continental crust that can generate significant volume of felsic melts. This in turn triggers emplacement of felsic intrusions that temporarily and spatially associate with the mafic dyke-like intrusions. The modeling results agree well with geological data from the Karelian Craton and provide possible explanation for the observed association of Paleoproterozoic mafic dykes and felsic intrusions which formed in a relatively short time interval (up to 20 Myrs) in the early stages of the supercontinent breakup.

Modelling chemical releases from fish farms: impact zones, dissolution time, and exposure probability

Abstract Tarpaulin bath treatments are used in open net-pen finfish aquaculture to combat parasitic infections, in particular sea lice. After treatment, the toxic wastewater is released directly into the ocean, potentially harming non-target species in the vicinity. We model the dispersion of wastewater chemicals using a high-resolution numerical ocean model. The results are used to estimate the impact area, impact range, dissolution time, and exposure probability for chemicals of arbitrary toxicity. The study area is a fish-farming intensive region on the Norwegian western coast. Simulations are performed at 61 different release dates, each on 16 locations. In our base case where the chemical is toxic at 1% of the treatment concentration, the release of a 16000 m³ wastewater plume traverses a median distance of 1.9 km before being completely dissolved. The median impacted area is 0.9 km² and the median dissolution time is 6.8 hours. These figures increase to 5.9 km, 7.0 km², and 21 hours, respectively, if the chemical is toxic at 0.1 % of the treatment concentration. Locations within fjords have slower dissolution rates and larger impact zones compared to exposed locations off the coast, especially during summer.

Generation Mechanism of Tidally-driven Whirlpools at A Narrow Strait in An Estuary

This study investigates the generation mechanism and influence of the whirlpools in the Naruto Strait on the surrounding marine environment using state-of-the-art high-resolution numerical ocean circulation modeling in a quadruple nested configuration. The Naruto whirlpools is recognized as an extraordinary seascape that the local governments and the citizens seek to register as a world natural heritage site. We found that the pronounced pressure gradient force associated with the meridional surface elevation difference was induced by a phase difference of two bifurcating major tidal waves. These waves originate from the Kitan Strait, and ultimately produce intense tidal currents at the Naruto Strait. One branch of the tidal waves propagates counter-clockwise along Awaji Island through the Akashi Strait, while the other occurs directly from the Kii Channel. As such, the whirlpool emerges as a large number of sub-mesoscale eddies, primarily due to the horizontal shear instability of tidal currents energized at the narrow topography between two headlands that extend into the strait. A dipole of overturning vertical circulations appears underneath the whirlpools with convergent downward transport at the strongest tidal current near the center of the strait; this causes efficient vertical mixing. This three-dimensional non-linear mixing promotes a time-averaged southeastward mass transport that extracts water and materials from the Harima-nada Sea into the Kii Channel.

Idealized Large-Eddy Simulations of Stratocumulus Advecting over Cold Water. Part I: Boundary Layer Decoupling

Abstract We explore the decoupling physics of a stratocumulus-topped boundary layer (STBL) moving over cooler water, a situation mimicking warm-air advection (WADV). We simulate an initially well-mixed STBL over a doubly periodic domain with the sea surface temperature decreasing linearly over time using the System for Atmospheric Modeling large-eddy model. Due to the surface cooling, the STBL becomes increasingly stably stratified, manifested as a near-surface temperature inversion topped by a well-mixed cloud-containing layer. Unlike the stably stratified STBL in cold-air advection (CADV) that is characterized by cumulus coupling, the stratocumulus deck in the WADV is unambiguously decoupled from the sea surface, manifested as weakly negative buoyancy flux throughout the subcloud layer. Without the influxes of buoyancy from the surface, the convective circulation in the well-mixed cloud-containing layer is driven by cloud-top radiative cooling. In such a regime, the downdrafts propel the circulation, in contrast to that in CADV regime for which the cumulus updrafts play a more determinant role. Such a contrast in convection regime explains the difference in many aspects of the STBLs including the entrainment rate, cloud homogeneity, vertical exchanges of heat and moisture, and lifetime of the stratocumulus deck, with the last being subject to a more thorough investigation in Part II. Finally, we investigate under what conditions a secondary stratus near the surface (or fog) can form in the WADV. We found that weaker subsidence favors the formation of fog whereas a more rapid surface cooling rate does not. Significant Statement The low-lying blanketlike clouds, called stratocumulus (Sc), reflect much incoming sunlight, substantially modulating Earth’s temperature. While much is known about how the Sc evolves when it moves over warmer water, few studies examine the opposite situation of Sc moving over colder water. We used a high-resolution numerical model to simulate such a case. When moving over cold water, the Sc becomes unambiguously decoupled from the water surface, distinctive from its warm counterpart in which the Sc interacts with the water surface via intermittent cauliflower-like clouds called cumulus clouds. Such decoupling influences many aspects of the Sc–sea surface system, which combine to alter the ability of the Sc to reflect sunlight, thereby influencing the climate. This work laid the foundation for future work that quantifies the contribution of such a decoupled Sc regime to Earth’s radiative budget and climate change.

Using rapid repeat SAR interferometry to improve hydrodynamic models of flood propagation in coastal wetlands

The propagation of tides and riverine floodwater in coastal wetlands is controlled by subtle topographic differences and a thick vegetation canopy. High-resolution numerical models have been used in recent years to simulate fluxes across wetlands. However, these models are based on sparse field data that can lead to unreliable results. Here, we utilize high spatial-resolution, rapid repeat interferometric data from the Uninhabited Aerial Vehicle Synthetic Aperture Radar (UAVSAR) to provide a synoptic measurement of sub-canopy water-level change resulting from tide propagation into wetlands. These data are used to constrain crucial model parameters and improve the performance and realism of simulations of the Wax Lake wetlands in coastal Louisiana (USA). A sensitivity analysis shows that the boundary condition of river discharge should be calibrated first, followed by iterative correction of terrain elevation specified originally by a Digital Terrain Model derived from LiDAR measurements. The calibration of bed friction becomes important only with the boundary and topography calibrated. With the model parameters calibrated, the overall Nash-Sutcliffe model efficiency for water-level change increases from 0.15 to 0.53 with the RMSE reduced by 26%. In areas with dense wetland grasses, the LiDAR signal is unable to reach the soil surface, but the L-band UAVSAR instrument detects changes in water levels that can be used to infer the true ground elevation. The high spatial resolution and repeat-acquisition frequency (minutes to hours) observations provided by UAVSAR represent a groundbreaking opportunity for a deeper understanding of the complex hydrodynamics of coastal wetlands.

Sea Surface Salinity Subfootprint Variability from a Global High-Resolution Model

Subfootprint variability (SFV) is variability at a spatial scale smaller than the footprint of a satellite, and it cannot be resolved by satellite observations. It is important to quantify and understand, as it contributes to the error budget for satellite data. The purpose of this study was to estimate the SFV for sea surface salinity (SSS) satellite observations. This was performed by using a high-resolution numerical model, a 1/48° version of the MITgcm simulation, from which one year of output has recently become available. SFV, defined as the weighted standard deviation of SSS within the satellite footprint, was computed from the model for a 2° × 2° grid of points for the one model year. We present maps of median SFV for 40 and 100 km footprint size, display histograms of its distribution for a range of footprint sizes and quantify its seasonality. At a 100 km (40 km) footprint size, SFV has a mode of 0.06 (0.04). It is found to vary strongly by location and season. It has larger values in western-boundary and eastern-equatorial regions, as well as in a few other areas. SFV has strong variability throughout the year, with the largest values generally being in the fall season. We also quantified the representation error, the degree of mismatch between random samples within a footprint and the footprint average. Our estimates of SFV and representation error can be used in understanding errors in the satellite observation of SSS.

Formulating Wave Overtopping at Vertical and Sloping Structures with Shallow Foreshores Using Deep-Water Wave Characteristics

The state-of-The-Art formulas for mean wave overtopping (q) assessment typically require wave conditions at the toe of the structure as input. However, for structures built either on land or in very shallow water, obtaining accurate estimates of wave height and period at the structure toe often proves difficult and requires the use of either physical modeling or high-resolution numerical wave models. Here, we follow Goda's method to establish an accurate prediction methodology for both vertical and sloping structures based entirely on deep-water characteristics-where the influence of the foreshore is captured by directly incorporating the foreshore slope and the relative water depth at the structure toe (htoe/Hm0,deep). Findings show that q decreases exponentially with htoe/Hm0,deep due to the decrease of the incident wave energy; however, the rate of reduction in q decreases for structures built on land or in extremely shallow water (htoe/Hm0,deep ≤ 0.1) due to the increased influence of wave-induced setup and infragravity waves-which act as long-period fluctuations in mean water level-generated by nonlinear wave transformation over the foreshore.

Comparison of the WRF and HARMONIE models ability for mountain wave warnings

Mountain lee waves usually involve aircraft icing and turbulence events. These weather phenomena, in turn, are a threat to aviation safety. For this reason, mountain lee waves are an interesting subject of study for the scientific community. This paper analyses several mountain lee waves events in the south-east of the Guadarrama mountain range, near the Adolfo Suarez Madrid-Barajas airport (Spain), using the Weather Research and Forecasting (WRF) and the HARMONIE-AROME high-resolution numerical models. For this work, simulated brightness temperature from the optimum WRF parametrization schemes and from the HARMONIE are validated using satellite observations to evaluate the performance of the models in reproducing the lenticular clouds associated to mountain lee waves. The brightness temperature probability density shows interesting differences between both models. Following, a mountain wave characterization is performed simulating some atmospheric variables (wind direction, wind speed, atmospheric stability, liquid water content and temperature) in several grid points located in the leeward, windward and over the summit of the mountains. The characterization results are compared for both numerical models and a decision tree is developed for each to forecast and warn the mountain lee waves, lenticular clouds and icing events with a 24 to 48 h lead time. These warnings are validated using several skill scores, revealing similar results for both models.

Quantitative Precipitation Forecasting Using an Improved Probability-Matching Method and Its Application to a Typhoon Event

This present study aims to explore how forecasters can quickly make accurate predictions by using various high-resolution model forecasts. Based on three high temporal-spatial resolution (3 km, hourly) numerical weather prediction models (CMA-MESO, CMA-GD, CMA-SH3) from the China Meteorological Administration (CMA), the hourly precipitation characteristics of three model within 24 h from March to September 2020 are discussed and integrated into a single, hourly, deterministic quantitative precipitation forecast (QPF) by making use of an improved weighted moving average probability-matching method (WPM). The results are as follows: (1) In non-rainstorm forecasts, CMA-MESO and CMA-GD have similar forecast abilities. However, in rainstorm forecasts, CMA-MESO has a notable advantage over the other two models. Thus, CMA-MESO is selected as a critical factor when participating in sensitivity experiments. (2) Compared with the traditional equal-weight probability-matching method (PM), the WPM improves the different grade QPF because it can effectively reduce rainfall pattern bias by making use of the weighted moving average (WMA). Additionally, the WPM threat score in rainstorm forecast similarly improved from 0.051 to 0.056, with a 9.8% increase relative to the PM. (3) The sensitivity experiments show that an optimal rainfall intensity score (WPM-best) can further improve the QPF and overcome all single models in both rainstorm and non-rainstorm forecasts, and the WPM-best has a rainstorm threat score skill of 0.062, with an increase of 21.6% compared with the PM. The performance of the WPM-best will be better if the precipitation intensity is stronger and the valid forecast periods is longer. It should be noted that there is no need to select models before using the WPM-best method, because WPM-best can give a very low weight to the less-skillful model in a more objective way. (4) The improved WPM method is also applied to investigate the heavy-rainfall case induced by typhoon Mekkhala (2020), where the improved WPM technique significantly improves rainstorm forecasting ability compared with a single model.