Model Initial Time Research Articles

Efficient Multivariate Time Series Anomaly Detection Through Transfer Learning for Large-Scale Software Systems

Timely anomaly detection of multivariate time series (MTS) is of vital importance for managing large-scale software systems. However, many deep learning-based MTS anomaly detection models require long-term MTS training data to achieve optimal performance, which often conflicts with the frequent pattern changes observed in software systems. Moreover, the training overhead of vast MTS in large-scale software systems is unacceptably high. To address these issues, we design OmniTransfer , a model-agnostic framework that combines weighted hierarchical agglomerative clustering with an adaptive transfer learning strategy, making many state-of-the-art (SOTA) MTS anomaly detection models efficient and effective. Extensive experiments using real-world data from a large web content service provider and a network operator show that OmniTransfer significantly reduces the model initialization time by 46.49% and the training cost by 74.51%, while maintaining high accuracy in detecting anomalies.

Improving Short-Term, Near-Surface Temperature Forecasts by Integrating Weather Pattern Information into Model Output Statistics

Abstract Dynamical numerical weather prediction has remarkably improved over the last decades. Yet postprocessing techniques are needed to calibrate forecasts which are based on statistical and machine learning techniques. With recent advances in the derivation of year-round, large-scale atmospheric circulations, or weather regimes, the question arises of whether this information can be valuable within forecast postprocessing methods. This paper investigates this by proposing a bias correction scheme to integrate the atmospheric circulation state derived from empirical orthogonal functions, referred to as weather patterns, for deterministic short-term, near-surface temperature forecasts based on least absolute shrinkage and selection operator (LASSO) regression. We propose a computational study which first evaluates different weather pattern definitions (spatial domain) to improve temperature forecasts in Europe. As a bias could be associated with the weather pattern at the model initialization time or at the realization time of the forecast, both variants are tested in this study. We show that forecasted weather patterns with the identical spatial domain as the forecast show best skill reaching mean-square-error skill improvements of up to 3% (day ahead) or 1% (week ahead). Only considering land surface improvements in Europe, improvements of 4%–6% for day-ahead forecasts and 1%–5% for week-ahead forecasts are observable. We believe that this study not only introduces a simple yet effective tool to reduce bias in temperature forecasts but also contributes to the active discussion of how valuable weather patterns are and how to use them within forecast calibration techniques.

Analysis of methods for assimilating fire perimeters into a coupled fire-atmosphere model

Correctly initializing the fire within coupled fire-atmosphere models is critical for producing accurate forecasts of meteorology near the fire, as well as the fire growth, and plume evolution. Improperly initializing the fire in a coupled fire-atmosphere model can introduce forecast errors that can impact wind circulations surrounding the fire and updrafts along the fire front. A well-constructed fire initialization process must be integrated within coupled fire-atmosphere models to ensure that the atmospheric component of the model does not become numerically unstable due to excessive heat fluxes released during the ignition, and that realistic fire-induced atmospheric circulations are established at the model initialization time. The primary objective of this study is to establish an effective fire initialization method in a coupled fire-atmosphere model, based on the analysis of the impact of the initialization procedure on the model’s ability to resolve fire-atmosphere circulations and fire growth. Here, we test three different fire initialization approaches leveraging the FireFlux II experimental fire, which provides a comprehensive suite of observations of the pyroconvective column, local micrometeorology, and fire characteristics. The two most effective fire initialization methods identified using the FireFlux II case study are then tested on the 380,000-acre Creek Fire, which burned across the central Sierra Nevada mountains during the 2020 Western U.S. wildfire season. For this case study, simulated pyroconvection and fire progression are evaluated using plume top height observations from MISR and airborne fire perimeter data, to assess the effectiveness of different initialization methods in the context of establishing pyroconvection and resolving the fire growth. The analyses of both the experimental fire simulation and the wildfire simulation indicate that the spin-up initialization method based on historical fire progression that masks out inactive fire regions provides the best results in terms of resolving the fire-induced vertical circulation and fire progression.

The Influence of Meteorology Initialization on Ozone Forecasting in the Great Lakes Region during MOOSE Study

This study investigates the influence of meteorology initialization on surface ozone prediction in the Great Lakes region using Canada’s operational air quality model (GEM-MACH) at a 2.5 km horizontal resolution. Two different initialization techniques are compared, and it is found that the four-dimensional incremental analysis updating (IAU) method yields improved model performance for surface ozone prediction. The IAU run shows better ozone regression line statistics (y = 0.7x + 14.9, R2 = 0.2) compared to the non-IAU run (y = 0.6x + 23.1, R2 = 0.1), with improved MB and NMB values (3.9 ppb and 8.9%, respectively) compared to the non-IAU run (4.1 ppb and 9.3%). Furthermore, analyzing ozone prediction sensitivity to model initialization time reveals that the 18z initialization leads to enhanced performance, particularly during high ozone exceedance days, with an improved regression slope of 0.9 compared to 0.7 for the 00z and 12z runs. The MB also improves to −0.2 ppb in the 18z run compared to −2.8 ppb and −3.9 ppb for the 00z and 12z runs, respectively. The analysis of meteorological fields reveals that the improved ozone predictions at 18z are linked to a more accurate representation of afternoon wind speed. This improvement enhances the transport of ozone, contributing to the overall improvement in ozone predictions.

The Forecast Skill of Tropical Cyclone Genesis in Two Global Ensembles

Abstract Tropical cyclone (TC) genesis forecasts during 2018–20 from two operational global ensemble prediction systems (EPSs) are evaluated over three basins in this study. The two ensembles are from the European Centre for Medium-Range Weather Forecasts (ECMWF-EPS) and the MetOffice in the United Kingdom (UKMO-EPS). The three basins include the northwest Pacific, northeast Pacific, and the North Atlantic. It is found that the ensemble members in each EPS show a good level of agreement in forecast skill, but their forecasts are complementary. Probability of detection (POD) can be doubled by taking all the member forecasts in the EPS into account. Even if an ensemble member does not make a hit forecast, it may predict the presence of cyclonic vortices. Statistically, a hit forecast has more nearby disturbance forecasts in the ensemble than a false alarm. Based on the above analysis, we grouped the nearby forecasts at each model initialization time to define ensemble genesis forecasts, and verified these forecasts to represent the performance of the ensemble system. The PODs are found to be more than twice that of the individual ensemble members at most lead times, which is about 59% and 38% at the 5-day lead time in UKMO-EPS and ECMWF-EPS, respectively; while the success ratios are smaller compared with that of the ensemble members. In addition, predictability differs in different basins, and genesis events in the North Atlantic basin are the most difficult to forecast in EPS, and its POD at the 5-day lead time is only 46% and 23% in UKMO-EPS and ECMWF-EPS, respectively. Significance Statement Operational forecasting of tropical cyclone (TC) genesis relies greatly on numerical models. Compared with deterministic forecasts, ensemble prediction systems (EPSs) can provide uncertainty information for forecasters. This study examined the predictability of TC genesis in two operational EPSs. We found that the forecasts of ensemble members complement each other, and the detection ratio of observed genesis will be doubled by considering the forecasts of all members, as multiple simulations conducted by the EPS partially reflect the inherent uncertainties of the genesis process. Successful forecasts are surrounded by more cyclonic vortices in the ensemble than false alarms, so the vortex information is used to group the nearby forecasts at each model initialization to define ensemble genesis forecasts when evaluating the ensemble performance. The results demonstrate that the global ensemble models can serve as a valuable reference for TC genesis forecasting.

Understanding the sensitivity of WRF hindcast of Beijing extreme rainfall of 21 July 2012 to microphysics and model initial time

The extreme rainfall during 21–22 July 2012 brought the heaviest rain to Beijing (190 mm) and Fangshan (460 mm) in six decades, which caused widespread damage in the city and its surroundings. To understand the evolution of spatio-temporal variability of precipitation and to quantify the inherent uncertainties in short-range weather forecasts of such an urban extreme precipitation event, we use an ensemble of high-resolution urban-aware simulations. In particular, we focus on the sensitivity of WRF hindcasts to microphysics parameterizations since reliable forecasting of extreme precipitation depends heavily on accurate descriptions of the processes of cloud microphysics. Three widely used microphysics schemes (Thompson, WSM6, and WDM6) were examined. Apart from the microphysics schemes, we also evaluated the performance of the simulations to four different model initial times (starting from 1200 UTC 19 July with 12 h-intervals) and two different boundary conditions (ERA-Interim and NCEP FNL). The results indicate that all three microphysics schemes underestimated total rainfall. The Thompson microphysics scheme, when initialized at the beginning of convection intensification (0000 UTC 20 July 2012), captures the intense rain areas. The WSM6 scheme captures the total amount of intense rainfall better than other schemes but with a spatial bias, and the WDM6 captures the location of the heavy precipitation but underpredicts the amount of precipitation compared to other schemes. The simulations starting at 0000 UTC 20 July 2012 perform better compared with other three other initialization times. The simulation using NCEP-FNL captures the location of maximum precipitation better compared with ERA-Interim. Overall, it is found that the results of these hindcasts are most sensitive to the time of initialization.

Track Uncertainty in High-Resolution HWRF Ensemble Forecasts of Hurricane Joaquin

Abstract Hurricane Joaquin (2015) was characterized by high track forecast uncertainty when it approached the Bahamas from 29 September 2015 to 1 October 2015, with 5-day track predictions ranging from landfall in the United States to east of Bermuda. The source of large track spread in Joaquin forecasts is investigated using an ensemble prediction system (EPS) based on the Hurricane Weather Research and Forecasting (HWRF) Model. For the first time, a high-resolution analysis of an HWRF-based EPS is performed to isolate the factors that control tropical cyclone (TC) track uncertainty. Differences in the synoptic-scale environment, the TC vortex structure, and the TC location are evaluated to understand the source of track forecast uncertainty associated with Joaquin, especially at later lead times when U.S. landfall was possible. EPS members that correctly propagated Joaquin into the central North Atlantic are compared with members that incorrectly predicted U.S. landfall. Joaquin track forecasts were highly dependent on the evolution of the environment, including weak atmospheric steering flow near the Bahamas and three synoptic-scale systems: a trough over North America, a ridge to the northeast of Joaquin, and an upper-tropospheric trough to the east of Joaquin. Differences in the steering flow were associated with perturbations of the synoptic-scale environment at the model initialization time. Ultimately, members that produced a more progressive midlatitude synoptic-scale pattern had reduced track errors. Joaquin track forecast uncertainty was not sensitive to the TC vortex structure or the initial TC position.

Impact of model improvements on 80 m wind speeds during the second Wind Forecast Improvement Project (WFIP2)

Abstract. During the second Wind Forecast Improvement Project (WFIP2; October 2015–March 2017, held in the Columbia River Gorge and Basin area of eastern Washington and Oregon states), several improvements to the parameterizations used in the High Resolution Rapid Refresh (HRRR – 3 km horizontal grid spacing) and the High Resolution Rapid Refresh Nest (HRRRNEST – 750 m horizontal grid spacing) numerical weather prediction (NWP) models were tested during four 6-week reforecast periods (one for each season). For these tests the models were run in control (CNT) and experimental (EXP) configurations, with the EXP configuration including all the improved parameterizations. The impacts of the experimental parameterizations on the forecast of 80 m wind speeds (wind turbine hub height) from the HRRR and HRRRNEST models are assessed, using observations collected by 19 sodars and three profiling lidars for comparison. Improvements due to the experimental physics (EXP vs. CNT runs) and those due to finer horizontal grid spacing (HRRRNEST vs. HRRR) and the combination of the two are compared, using standard bulk statistics such as mean absolute error (MAE) and mean bias error (bias). On average, the HRRR 80 m wind speed MAE is reduced by 3 %–4 % due to the experimental physics. The impact of the finer horizontal grid spacing in the CNT runs also shows a positive improvement of 5 % on MAE, which is particularly large at nighttime and during the morning transition. Lastly, the combined impact of the experimental physics and finer horizontal grid spacing produces larger improvements in the 80 m wind speed MAE, up to 7 %–8 %. The improvements are evaluated as a function of the model's initialization time, forecast horizon, time of the day, season of the year, site elevation, and meteorological phenomena. Causes of model weaknesses are identified. Finally, bias correction methods are applied to the 80 m wind speed model outputs to measure their impact on the improvements due to the removal of the systematic component of the errors.

Sensitivity of Idealized Hurricane Intensity and Structures under Varying Background Flows and Initial Vortex Intensities to Different Vertical Resolutions in HWRF

Abstract A series of 5-day numerical simulations of idealized hurricane vortices under the influence of different background flows is performed by varying vertical grid resolution (VGR) in different portions of the atmosphere with the operational version of the Hurricane Weather Research and Forecasting Model in order to study the sensitivity of hurricane intensity forecasts to different distributions of VGR. Increasing VGR from 21 to 43 levels produces stronger hurricanes, whereas increasing it further to 64 levels does not intensify the storms further, but intensity fluctuations are much reduced. Moreover, increasing the lower-level VGRs generates stronger storms, but the opposite is true for increased upper-level VGRs. On average, adding mean flow increases intensity fluctuations and variability (between the strongest and weakest hurricanes), whereas adding vertical wind shear (VWS) delays hurricane intensification and then causes more rapid growth in intensity variability. The stronger the VWS, the larger intensity variability and bifurcation rate occur at later stages. These intensity differences are found to be closely related to inner-core structural changes, and they are attributable to how much latent heat could be released in higher-VGR layers, followed by how much moisture content in nearby layers is converged. Hurricane intensity with higher VGRs is shown to be much less sensitive to varying background flows, and stronger hurricane vortices at the model initial time are less sensitive to the vertical distribution of VGR; the opposite is true for relatively uniform VGRs or weaker hurricane vortices. Results reveal that higher VGRs with a near-parabolic or Ω shape tend to produce smoother intensity variations and more typical inner-core structures.

Influence of orography on precipitation amount and distribution in NW Greece; A case study

In this work the impact of topography on the distribution and the amount of precipitation in Epirus, NW Greece is investigated, during the cold period of the year. A precipitation event with typical wintertime characteristics is considered. In the selected case, extended and heavy precipitation over Epirus is caused by a low pressure system centred northwest of Epirus and moving southeastwards, generating southerly winds over the Ionian Sea. The case is simulated by applying the numerical meteorological model MM5 in a high resolution grid (2×2km). Various sensitivity tests are performed, searching for the model initialization time and the convective parameterization scheme (CPS) leading to the best simulation of the observed event. Then, the model is applied again incorporating modified data of topography, in order to study the differentiations in the fields of the most important meteorological parameters, which are directly related with precipitation. The topography is modified with two ways: either by reducing the height of the mountains (type RT) or by elevating the terrain height (type ET). The results show that in the RT type, the precipitation field appears more homogeneous, with weaker maxima on the windward sides and above the mountain ridges. As a consequence, less rain falls in these areas and humid air passes easier to the lee sides reducing rain shadows. In the ET type, rain increases in the windward coastal zone of Epirus while it is reduced in the inland areas where less humid air is available. Generally, the modification of topography changes the flows near the surface, causing a displacement of the convergence zones and thus a new distribution of vertical velocities, which, combined with the new water vapour field, has a direct impact on precipitation.

Variational assimilation of HF radar surface currents in a coastal ocean model off Oregon

The impact of assimilation of sea surface velocity fields observed by a set of high-frequency (HF) radars is studied using a three-dimensional ocean circulation model configured along the Oregon coast. The study period is June–July 2008 featuring upwelling and separation of the coastal currents into the adjacent interior ocean. The nonlinear model is based on the Regional Ocean Modeling System (ROMS) and the data assimilation (DA) component on the AVRORA system utilizing the representer-based variational algorithm. Assimilation proceeds in a series of 3-day windows, providing an analysis solution in each window and a 3-day forecast into the next window. Experiments with two different initial condition error covariances are compared (one is dynamically balanced, based on the linearized equation of state, temperature-salinity relation, and geostrophic and thermal wind balance relations and the other is multivariate unbalanced). While the assimilation impact is statistically better in the case of the balanced covariance, the case with the unbalanced covariance also provides sensible improvement in terms of surface velocity and sea surface temperature (SST) model-data forecast statistics. The analysis of representer functions shows that even if the initial condition error covariance is unbalanced, the correction fields at the model initial time are partially balanced after each dynamical field is smoothed independently, due to inherent dynamical properties of the adjoint model. Assimilation of the HF radar surface currents improves not only surface velocity forecasts, but also geometry of the upwelling SST front and the sea surface height (SSH) slope near the coast, as verified against unassimilated satellite SSH and SST data. The assimilation also alters the latitudinal distribution of the time-averaged offshore transport. Combined HF radar velocity and other observations, e.g., altimetry, is needed to better constrain surface geostrophic currents in the entire model domain, including the area not covered by the HF radar data.

Hydrometeorological Modeling Study of Tropical Cyclone Phet in the Arabian Sea in 2010

Tropical cyclone Phet is the second strongest tropical cyclone ever recorded in the Arabian Sea. Phet made landfall in the northeast mountainous area of Oman in early morning on 4 June in 2010, causing a breaking record rainfall in this arid region of 488 mm/48 h. The cyclone heavy rainfall triggered flash floods causing enormous losses in lives and infrastructure in northeast Oman. The state of the art Advanced Research WRF model is used to study the atmospheric circulation and to reproduce the heavy rainfall over Oman. Three one-way nested domains with 32 vertical layers with terrain following sigma coordinate are used to setup eight numerical experiments aiming to investigate the effect of initialization time, horizontal grid resolution and terrain elevations on reproducing the cyclone track, intensity and heavy rainfall. Simulation results show negligible effect of model initialization time on cyclone track, intensity and rainfall. In contrast, the orographic effect played a substantial role in rainfall simulation over northeast Oman. The heavy rainfall was a combination of the cyclone circulation effect and the orographic lifting in the mountains. The northeasterly cyclone moist-warm wind was lifted in the Omani mountains releasing its potential energy and enhancing further thermal convection. The numerical experiment with the highest terrain elevation (RUN3.3-C) resulted in overestimation of observed rainfall due to the enhanced topographic lifting of the saturated cyclone wind. Experiment with similar horizontal grid resolution but smoother terrain elevation (RUN3.3-TER) resulted in much less rainfall amount comparable to the observed values. The increased precipitation in RUN3.3-C is due to the increase in the rain- water and cloud water and graupel of the explicit moisture scheme.

Experiments of Hurricane Initialization with Airborne Doppler Radar Data for the Advanced Research Hurricane WRF (AHW) Model

Abstract Initialization of the hurricane vortex in weather prediction models is vital to intensity forecasts out to at least 48 h. Airborne Doppler radar (ADR) data have sufficiently high horizontal and vertical resolution to resolve the hurricane vortex and its imbedded structures but have not been extensively used in hurricane initialization. Using the Weather Research and Forecasting (WRF) three-dimensional variational data assimilation (3DVAR) system, the ADR data are assimilated to recover the hurricane vortex dynamic and thermodynamic structures at the WRF model initial time. The impact of the ADR data on three hurricanes, Jeanne (2004), Katrina (2005) and Rita (2005), are examined during their rapid intensification and subsequent weakening periods before landfall. With the ADR wind data assimilated, the three-dimensional winds in the hurricane vortex become stronger and the maximum 10-m winds agree better with independent estimates from best-track data than without ADR data assimilation. Through the multivariate incremental structure in WRF 3DVAR analysis, the central sea level pressures (CSLPs) for the three hurricanes are lower in response to the stronger vortex at initialization. The size and inner-core structure of each vortex are adjusted closer to observations of these attributes. Addition of reflectivity data in assimilation produces cloud water and rainwater analyses in the initial vortex. The temperature and moisture are also better represented in the hurricane initialization. Forty-eight-hour forecasts are conducted to evaluate the impact of ADR data using the Advanced Research Hurricane WRF (AHW), a derivative of the Advanced Research WRF (ARW) model. Assimilation of ADR data improves the hurricane-intensity forecasts. Vortex asymmetries, size, and rainbands are also simulated better. Hurricane initialization with ADR data is quite promising toward reducing intensity forecast errors at modest computational expense.

Impact of GPS Radio Occultation Refractivity Soundings on a Simulation of Typhoon Bilis (2006) upon Landfall

Typhoon Bilis which struck Taiwan in July 2006 was chosen to assess the potential impact of GPS radio occultation (RO) refractivity soundings on numerical simulation using the WRF model. We found that this case elucidates the impact of the limited GPS RO soundings on typhoon prediction due to their favorable locations. In addition, on top of available precipitable water (PW) and near-surface wind speed from S SM/I data, we have also explored their combined impacts on model prediction. The two GPS RO soundings available from FORMOSAT-3 near the model initial time are assimilated using the nonlocal operator in WRF 3DVAR. More contributions after the assimilation are found in the moisture increments (up to -1.2 g kg^(-1)) surrounding the corresponding occultation positions to the east of the typhoon vortex and southwest of the subtropical high, with dryness occurring in the lower troposphere. The assimilation of PW induces much larger and wider moisture increments to the west of the typhoon center. The ingestion of near-surface wind speeds produces more pronounced temperature and wind increments at lower levels. The S SM/I data improve the predictions of typhoon intensity and intense rainfall over Taiwan, especially for the later simulation time, but generally have a less impact on track prediction as compared to the GPS RO data. Assimilation of the two GPS RO soundings leads to a marked improvement on the track that otherwise will be southward biased prior to landfall and also enhances the positive impact when the SSMII data have been assimilated as well. An improvement using the GPS RO data is also found for daily rainfall but only for larger thresholds at later times when the bias in track in the control experiment is remedied. In general, the combined assimilation of both GPS RO and SSM/I data has shown the most promising performances. Finally, the sensitivity tests indicate that the closer GPS RO sounding to the east of the typhoon plays a more crucial role in the positive impact on track prediction.

An Examination of Precipitation in Observations and Model Forecasts during NAME with Emphasis on the Diurnal Cycle

Abstract Summertime rainfall over the United States and Mexico is examined and is compared with forecasts from operational numerical prediction models. In particular, the distribution of rainfall amounts is examined and the diurnal cycle of rainfall is investigated and compared with the model forecasts. This study focuses on a 35-day period (12 July–15 August 2004) that occurred amid the North American Monsoon Experiment (NAME) field campaign. Three-hour precipitation forecasts from the numerical models were validated against satellite-derived estimates of rainfall that were adjusted by daily rain gauge data to remove bias from the remotely sensed estimates. The model forecasts that are evaluated are for the 36–60-h period after the model initial run time so that the effects of updated observational data are reduced substantially and a more direct evaluation of the model precipitation parameterization can be accomplished. The main findings of this study show that the effective spatial resolution of the model-generated precipitation is considerably more coarse than the native model resolution. On a national scale, the models overforecast the frequency of rainfall events in the 1–75 mm day−1 range and underforecast heavy events (>85 mm day−1). The models also have a diurnal cycle that peaks 3–6 h earlier than is observed over portions of the eastern United States and the NAME tier-1 region. Time series and harmonic analysis are used to identify where the models perform well and poorly in characterizing the amplitude and phase of the diurnal cycle of precipitation.

A Numerical Study of the 1996 Saguenay Flood Cyclone: Effect of Assimilation of Precipitation Data on Quantitative Precipitation Forecasts

Abstract A one-dimensional variational (1DVAR) technique is applied to assimilate rain gauge precipitation data to extend the predictability of the Saguenay flood cyclone associated with a trough-merger event on 19–21 July 1996 in the Saguenay-Lac-Saint-Jean region of Quebec, Canada. Two 60-h simulations initialized at 0000 UTC 19 July were performed with the Canadian Mesoscale Compressible Community (MC2) model. The control (CTL) and NCEP simulations were initialized with the enhanced temperature and moisture profiles obtained from the 1DVAR scheme and the NCEP reanalysis data, respectively. Compared to observations, the CTL simulation reasonably reproduced the observed mass and wind fields and showed a marked improvement in the threat scores for heavy precipitation. The CTL run captured the observed spatial and temporal distribution of precipitation but overpredicted the area of precipitation. Sensitivity experiments showed that the threat (bias) scores are less (somewhat) sensitive to the specification of the observation error of the precipitation data. Of the four precipitation systems present at model initial time, the systems in the vicinity of the southern trough had the biggest impact on the threat score. Potential vorticity diagnostics of the CTL simulation suggested that the initial temperature and moisture field near the southern trough decreased the condensational heating relative to NCEP. This resulted in a stronger zonal wind component in the upper levels associated with the southern trough in CTL that retarded the eastward propagation of the northern trough, resulting in a correct placement of the surface precipitation and an improvement in the threat scores relative to NCEP.

NWP Cloud Initialization Using GOES Sounder Data and Improved Modeling of Nonprecipitating Clouds

Cloud-top pressures and effective cloud amounts, derived from Geostationary Operational Environmental Satellite (GOES) sounders, are used to specify a three-dimensional mesoscale cloud field for numerical weather prediction. A bulk cloud parameterization scheme based on relative humidity is introduced into the forecast model to efficiently use the satellite sounder observations. The cloud initialization and bulk cloud parameterization are tested in the Cooperative Institute for Meteorological Satellite Studies Regional Assimilation System for summer and winter periods. Comparisons between cloud forecasts and GOES cloud observations for four model configurations are presented. It is shown that the use of the initialization improved the areal coverage of the nonprecipitating clouds early in the forecast period, which translated to an improvement of the precipitation equitable threat score for the 24-h forecast (when the model initial time was 1200 UTC) and the 36-h forecast (when the model initial time was 0000 UTC).

A Simple Technique for Using Radar Data in the Dynamic Initialization of a Mesoscale Model

Abstract A simple technique for using radar reflectivity to improve model initialization is presented. Unlike previous techniques, the scheme described here does not infer rain rates and heating profiles from assumed relationships between remotely sensed variables and precipitation rates. Rather, the radar data are only used to tell the model when and where deep moist convection is occurring. This information is then used to activate the model’s convective parameterization scheme in the grid elements where convection is observed. This approach has the advantage that the convective precipitation rates and heating profiles generated by the convective parameterization are compatible with the local (grid element) environment. The premise is that if convection is forced to develop when and where it is observed during a data assimilation period, convectively forced modifications to the environment will be in the correct locations at the model initial forecast time and the resulting forecast will be more accurat...

An Objective Method to Modify Numerical Model Forecasts with Newly Given Weather Data Using an Artificial Neural Network

An objective method of forecasting precipitation coverage with a neural network is presented. This method uses as predictors all available data at local weather stations including both numerical model results and weather data obtained later than the model initial time, which sometimes contradict each other and hence have to be handled subjectively by well-experienced forecasters. Since the method gives an objective and also realistic forecast of areal precipitation coverage, its skill scores are better than those of the persistence forecast (after 3 h), the linear regression forecasts, and numerical model precipitation prediction.

Numerical Simulation of an Intense Squall Line during 10–11 June 1985 PRE-STORM. Part I: Model Verification

Abstract A 24-h nested-grid simulation of an intense squall line during the 1985 PRE-STORM experiment is presented using an improved version of the Pennsylvania State University/National Center for Atmospheric Research three-dimensional mesoscale model. Although the model is initialized at 1200 UTC 10 June 1985 with conventional meteorological observations, it reproduces remarkably well many observed meso-β scale features that are analyzed from the high-resolution network data. These include 1) the generation of two areas of deep convection at the model initial time; 2) the timing of the initiation of the squall line along a surface front 9 h into the model integration; 3) the development of several convective bands at 2100 UTC; 4) the rapid intensification and rapid dissipation processes of the squall line as it entered and moved out of the network, respectively; 5) the generation of a presquall mesolow, a squall-induced mesohigh and a wake low as well as corresponding multiple surface convergence-diverg...