Search Results
You are looking at 1 - 10 of 75 items for
- Author or Editor: William A. Gallus Jr. x
- Refine by Access: All Content x
Abstract
A 10-km-grid-spacing version of NCEP's Eta Model was used to simulate 11 warm-season convective systems occurring over the U.S. upper midwest. Quantitative precipitation forecasts (QPFs) from the model valid for 6-h periods were verified using 4-km-grid-spacing stage-IV precipitation estimates. Verification first was performed on the model's 10-km grid by areally averaging the 4-km observations onto the model grid. To investigate and quantify the impact of the verification grid-box size on some standard skill scores, verification was also performed by averaging the 10-km model forecasts onto 30-km grid boxes and then areally averaging the observations onto the same 30-km grid. As a final test of the impact of the verifying grid-box size, the same 11 events were simulated with a 30-km version of the Eta Model, with verification then being performed on this 30-km grid. For all cases in both the 10- and 30-km versions of the model, 12 variations of the model were used, with variations involving either (i) modifications to the initial conditions to better represent mesoscale features present at the initialization time or (ii) changes in moist physics. Equitable threat scores (ETSs) increased when verification occurred on a coarser grid, whether the coarser grid was created by averaging the 10-km model results or was that used in the 30-km model runs. This result suggests that it may be difficult to show improved skill scores as model resolution improves if the verification is performed on the model's own increasingly fine grid. It should be noted, however, that the use of different verification resolutions does not change the general impacts on ETSs of variations in model physics or initial conditions. The sensitivity of ETSs to verifying grid-box size does, however, vary somewhat between model variants using different model moist-physics formulations or initialization procedures.
Abstract
A 10-km-grid-spacing version of NCEP's Eta Model was used to simulate 11 warm-season convective systems occurring over the U.S. upper midwest. Quantitative precipitation forecasts (QPFs) from the model valid for 6-h periods were verified using 4-km-grid-spacing stage-IV precipitation estimates. Verification first was performed on the model's 10-km grid by areally averaging the 4-km observations onto the model grid. To investigate and quantify the impact of the verification grid-box size on some standard skill scores, verification was also performed by averaging the 10-km model forecasts onto 30-km grid boxes and then areally averaging the observations onto the same 30-km grid. As a final test of the impact of the verifying grid-box size, the same 11 events were simulated with a 30-km version of the Eta Model, with verification then being performed on this 30-km grid. For all cases in both the 10- and 30-km versions of the model, 12 variations of the model were used, with variations involving either (i) modifications to the initial conditions to better represent mesoscale features present at the initialization time or (ii) changes in moist physics. Equitable threat scores (ETSs) increased when verification occurred on a coarser grid, whether the coarser grid was created by averaging the 10-km model results or was that used in the 30-km model runs. This result suggests that it may be difficult to show improved skill scores as model resolution improves if the verification is performed on the model's own increasingly fine grid. It should be noted, however, that the use of different verification resolutions does not change the general impacts on ETSs of variations in model physics or initial conditions. The sensitivity of ETSs to verifying grid-box size does, however, vary somewhat between model variants using different model moist-physics formulations or initialization procedures.
Abstract
Both the Method for Object-based Diagnostic Evaluation (MODE) and contiguous rain area (CRA) object-based verification techniques have been used to analyze precipitation forecasts from two sets of ensembles to determine if spread-skill behavior observed using traditional measures can be seen in the object parameters. One set consisted of two eight-member Weather Research and Forecasting (WRF) model ensembles: one having mixed physics and dynamics with unperturbed initial and lateral boundary conditions (Phys) and another using common physics and a dynamic core but with perturbed initial and lateral boundary conditions (IC/LBC). Traditional measures found that spread grows much faster in IC/LBC than in Phys so that after roughly 24 h, better skill and spread are found in IC/LBC. These measures also reflected a strong diurnal signal of precipitation. The other set of ensembles included five members of a 4-km grid-spacing WRF ensemble (ENS4) and five members of a 20-km WRF ensemble (ENS20). Traditional measures suggested that the diurnal signal was better in ENS4 and spread increased more rapidly than in ENS20.
Standard deviations (SDs) of four object parameters computed for the first set of ensembles using MODE and CRA showed the trend of enhanced spread growth in IC/LBC compared to Phys that had been observed in traditional measures, with the areal coverage of precipitation exhibiting the greatest growth in spread with time. The two techniques did not produce identical results; although, they did show the same general trends. A diurnal signal could be seen in the SDs of all parameters, especially rain rate, volume, and areal coverage. MODE results also found evidence of a diurnal signal and faster growth of spread in object parameters in ENS4 than in ENS20.
Some forecasting approaches based on MODE and CRA output are also demonstrated. Forecasts based on averages of object parameters from each ensemble member were more skillful than forecasts based on MODE or CRA applied to an ensemble mean computed using the probability matching technique for areal coverage and volume, but differences in the two techniques were less pronounced for rain rate and displacement. The use of a probability threshold to define objects was also shown to be a valid forecasting approach with MODE.
Abstract
Both the Method for Object-based Diagnostic Evaluation (MODE) and contiguous rain area (CRA) object-based verification techniques have been used to analyze precipitation forecasts from two sets of ensembles to determine if spread-skill behavior observed using traditional measures can be seen in the object parameters. One set consisted of two eight-member Weather Research and Forecasting (WRF) model ensembles: one having mixed physics and dynamics with unperturbed initial and lateral boundary conditions (Phys) and another using common physics and a dynamic core but with perturbed initial and lateral boundary conditions (IC/LBC). Traditional measures found that spread grows much faster in IC/LBC than in Phys so that after roughly 24 h, better skill and spread are found in IC/LBC. These measures also reflected a strong diurnal signal of precipitation. The other set of ensembles included five members of a 4-km grid-spacing WRF ensemble (ENS4) and five members of a 20-km WRF ensemble (ENS20). Traditional measures suggested that the diurnal signal was better in ENS4 and spread increased more rapidly than in ENS20.
Standard deviations (SDs) of four object parameters computed for the first set of ensembles using MODE and CRA showed the trend of enhanced spread growth in IC/LBC compared to Phys that had been observed in traditional measures, with the areal coverage of precipitation exhibiting the greatest growth in spread with time. The two techniques did not produce identical results; although, they did show the same general trends. A diurnal signal could be seen in the SDs of all parameters, especially rain rate, volume, and areal coverage. MODE results also found evidence of a diurnal signal and faster growth of spread in object parameters in ENS4 than in ENS20.
Some forecasting approaches based on MODE and CRA output are also demonstrated. Forecasts based on averages of object parameters from each ensemble member were more skillful than forecasts based on MODE or CRA applied to an ensemble mean computed using the probability matching technique for areal coverage and volume, but differences in the two techniques were less pronounced for rain rate and displacement. The use of a probability threshold to define objects was also shown to be a valid forecasting approach with MODE.
Abstract
A two-dimensional cloud model is used to investigate whether microphysical processes alone within the stratiform rain regions of mesoscale convection systems can induce strong descent and intense surface wake lows accompanying such systems. Idealized simulations are run with a domain that captures the back edge of the stratiform rain region. A simplified microphysical field, representing snow alone, is prescribed within the stratiform cloud to produce radar reflectivities similar to observations. When the prescribed snow field is assumed time-independent, strong subsidence develops but does not induce an intense wake low since microphysical cooling strongly opposes adiabatic warming. Simply increasing snow quantities, although resulting in heavier rain rates and stronger subsidence, does not produce significant pressure falls. However, when precipitation rates are prescribed to decrease with time as might occur with collapsing precipitation cores, subsidence induces greater pressure falls, and a tighter pressure gradient near the wake low, in better agreement with observations.
Abstract
A two-dimensional cloud model is used to investigate whether microphysical processes alone within the stratiform rain regions of mesoscale convection systems can induce strong descent and intense surface wake lows accompanying such systems. Idealized simulations are run with a domain that captures the back edge of the stratiform rain region. A simplified microphysical field, representing snow alone, is prescribed within the stratiform cloud to produce radar reflectivities similar to observations. When the prescribed snow field is assumed time-independent, strong subsidence develops but does not induce an intense wake low since microphysical cooling strongly opposes adiabatic warming. Simply increasing snow quantities, although resulting in heavier rain rates and stronger subsidence, does not produce significant pressure falls. However, when precipitation rates are prescribed to decrease with time as might occur with collapsing precipitation cores, subsidence induces greater pressure falls, and a tighter pressure gradient near the wake low, in better agreement with observations.
Abstract
Simulations were performed using the Eta Model with its eta vertical coordinate and stepwise treatment of terrain, and with a substitution of the terrain-following sigma vertical coordinate to investigate the impact of step orography on flow near high mountains. Two different cases were simulated: (i) a downslope windstorm along the Front Range of the Rocky Mountains, and (ii) stably stratified flow blocked by high mountains in Taiwan. Flow separation on the lee side of the mountains, previously shown to occur in idealized two-dimensional Eta simulations, was also apparent in these real data cases, even for the downslope wind event. The step orography resulted in a substantial underestimate of wind speeds to the lee of the Rockies during the windstorm. Near the surface, both the eta and sigma simulations of the Taiwan blocking event were comparable. For both events, the use of step orography resulted in much weaker mountain waves than occurred when the sigma vertical coordinate was used. Localized vertical velocity perturbations associated directly with the step corners were minor for these cases.
Abstract
Simulations were performed using the Eta Model with its eta vertical coordinate and stepwise treatment of terrain, and with a substitution of the terrain-following sigma vertical coordinate to investigate the impact of step orography on flow near high mountains. Two different cases were simulated: (i) a downslope windstorm along the Front Range of the Rocky Mountains, and (ii) stably stratified flow blocked by high mountains in Taiwan. Flow separation on the lee side of the mountains, previously shown to occur in idealized two-dimensional Eta simulations, was also apparent in these real data cases, even for the downslope wind event. The step orography resulted in a substantial underestimate of wind speeds to the lee of the Rockies during the windstorm. Near the surface, both the eta and sigma simulations of the Taiwan blocking event were comparable. For both events, the use of step orography resulted in much weaker mountain waves than occurred when the sigma vertical coordinate was used. Localized vertical velocity perturbations associated directly with the step corners were minor for these cases.
Abstract
A versatile workstation version of the NCEP Eta Model is used to simulate three excessive precipitation episodes in the central United States. These events all resulted in damaging flash flooding and include 16–17 June 1996 in the upper Midwest, 17 July 1996 in western Iowa, and 27 May 1997 in Texas. The episodes reflect a wide range of meteorological situations: (i) a warm core cyclone in June 1996 generated a meso-β-scale region of excessive rainfall from echo training in its warm sector while producing excessive overrunning rainfall to the north of its warm front, (ii) a mesoscale convective complex in July 1996 produced excessive rainfall, and (iii) tornadic thunderstorms in May 1997 resulted in small-scale excessive rains.
Model sensitivity to horizontal resolution is investigated using a range of horizontal resolutions comparable to those used in operational and quasi-operational forecasting models. Sensitivity tests are also performed using both the Betts–Miller–Janjic convective scheme (operational at NCEP in 1998) and the Kain–Fritsch scheme. Variations in predicted peak precipitation as resolution is refined are found to be highly case dependent, suggesting forecaster interpretation of increasingly higher resolution model quantitative precipitation forecast (QPF) information will not be straightforward. In addition, precipitation forecasts and QPF response to changing resolution are both found to vary significantly with choice of convective parameterization.
Abstract
A versatile workstation version of the NCEP Eta Model is used to simulate three excessive precipitation episodes in the central United States. These events all resulted in damaging flash flooding and include 16–17 June 1996 in the upper Midwest, 17 July 1996 in western Iowa, and 27 May 1997 in Texas. The episodes reflect a wide range of meteorological situations: (i) a warm core cyclone in June 1996 generated a meso-β-scale region of excessive rainfall from echo training in its warm sector while producing excessive overrunning rainfall to the north of its warm front, (ii) a mesoscale convective complex in July 1996 produced excessive rainfall, and (iii) tornadic thunderstorms in May 1997 resulted in small-scale excessive rains.
Model sensitivity to horizontal resolution is investigated using a range of horizontal resolutions comparable to those used in operational and quasi-operational forecasting models. Sensitivity tests are also performed using both the Betts–Miller–Janjic convective scheme (operational at NCEP in 1998) and the Kain–Fritsch scheme. Variations in predicted peak precipitation as resolution is refined are found to be highly case dependent, suggesting forecaster interpretation of increasingly higher resolution model quantitative precipitation forecast (QPF) information will not be straightforward. In addition, precipitation forecasts and QPF response to changing resolution are both found to vary significantly with choice of convective parameterization.
Abstract
Bow echo structures, a subset of mesoscale convective systems (MCSs), are often poorly forecast within deterministic numerical weather prediction model simulations. Among other things, this may be due to the inherent low predictability associated with bow echoes, deficient initial conditions (ICs), and inadequate parameterization schemes. Four different ensemble configurations assessed the sensitivity of the MCSs’ simulated reflectivity and radius of curvature to the following: perturbations in initial and lateral boundary conditions using a global dataset, different microphysical schemes, a stochastic kinetic energy backscatter (SKEB) scheme, and a mix of the previous two. One case is poorly simulated no matter which IC dataset or microphysical parameterization is used. In the other case, almost all simulations reproduce a bow echo. When the IC dataset and microphysical parameterization is fixed within a SKEB ensemble, ensemble uncertainty is smaller. However, while differences in the location and timing of the MCS are reduced, variations in convective mode remain substantial. Results suggest the MCS’s positioning is influenced primarily by ICs, but its mode is most sensitive to the model error uncertainty. Hence, correct estimation of model error uncertainty on the storm scale is crucial for adequate spread and the probabilistic forecast of convective events.
Abstract
Bow echo structures, a subset of mesoscale convective systems (MCSs), are often poorly forecast within deterministic numerical weather prediction model simulations. Among other things, this may be due to the inherent low predictability associated with bow echoes, deficient initial conditions (ICs), and inadequate parameterization schemes. Four different ensemble configurations assessed the sensitivity of the MCSs’ simulated reflectivity and radius of curvature to the following: perturbations in initial and lateral boundary conditions using a global dataset, different microphysical schemes, a stochastic kinetic energy backscatter (SKEB) scheme, and a mix of the previous two. One case is poorly simulated no matter which IC dataset or microphysical parameterization is used. In the other case, almost all simulations reproduce a bow echo. When the IC dataset and microphysical parameterization is fixed within a SKEB ensemble, ensemble uncertainty is smaller. However, while differences in the location and timing of the MCS are reduced, variations in convective mode remain substantial. Results suggest the MCS’s positioning is influenced primarily by ICs, but its mode is most sensitive to the model error uncertainty. Hence, correct estimation of model error uncertainty on the storm scale is crucial for adequate spread and the probabilistic forecast of convective events.
Abstract
Nocturnal bow echoes can produce wind damage, even in situations where elevated convection occurs. Accurate forecasts of wind potential tend to be more challenging for operational forecasters than for daytime bows because of incomplete understanding of how elevated convection interacts with the stable boundary layer. The present study compares the differences in warm-season, nocturnal bow echo environments in which high intensity [>70 kt (1 kt ≈ 0.51 m s−1)] severe winds (HS), low intensity (50–55 kt) severe winds (LS), and nonsevere winds (NS) occurred. Using a sample of 132 events from 2010 to 2018, 43 forecast parameters from the SPC mesoanalysis system were examined over a 120 km × 120 km region centered on the strongest storm report or most pronounced bowing convective segment. Severe composite parameters are found to be among the best discriminators between all severity types, especially derecho composite parameter (DCP) and significant tornado parameter (STP). Shear parameters are significant discriminators only between severe and nonsevere cases, while convective available potential energy (CAPE) parameters are significant discriminators only between HS and LS/NS bow echoes. Convective inhibition (CIN) is among the worst discriminators for all severity types. The parameters providing the most predictive skill for HS bow echoes are STP and most unstable CAPE, and for LS bow echoes these are the V wind component at best CAPE (VMXP) level, STP, and the supercell composite parameter. Combinations of two parameters are shown to improve forecasting skill further, with the combination of surface-based CAPE and 0–6-km U shear component, and DCP and VMXP, providing the most skillful HS and LS forecasts, respectively.
Abstract
Nocturnal bow echoes can produce wind damage, even in situations where elevated convection occurs. Accurate forecasts of wind potential tend to be more challenging for operational forecasters than for daytime bows because of incomplete understanding of how elevated convection interacts with the stable boundary layer. The present study compares the differences in warm-season, nocturnal bow echo environments in which high intensity [>70 kt (1 kt ≈ 0.51 m s−1)] severe winds (HS), low intensity (50–55 kt) severe winds (LS), and nonsevere winds (NS) occurred. Using a sample of 132 events from 2010 to 2018, 43 forecast parameters from the SPC mesoanalysis system were examined over a 120 km × 120 km region centered on the strongest storm report or most pronounced bowing convective segment. Severe composite parameters are found to be among the best discriminators between all severity types, especially derecho composite parameter (DCP) and significant tornado parameter (STP). Shear parameters are significant discriminators only between severe and nonsevere cases, while convective available potential energy (CAPE) parameters are significant discriminators only between HS and LS/NS bow echoes. Convective inhibition (CIN) is among the worst discriminators for all severity types. The parameters providing the most predictive skill for HS bow echoes are STP and most unstable CAPE, and for LS bow echoes these are the V wind component at best CAPE (VMXP) level, STP, and the supercell composite parameter. Combinations of two parameters are shown to improve forecasting skill further, with the combination of surface-based CAPE and 0–6-km U shear component, and DCP and VMXP, providing the most skillful HS and LS forecasts, respectively.
Abstract
Nocturnal mesoscale convective systems (MCSs) are important phenomena because of their contributions to warm-season precipitation and association with severe hazards. Past studies have shown that their morphology remains poorly forecast in current convection-allowing models operating at 3–4-km horizontal grid spacing. A total of 10 MCS cases occurring in weakly forced environments were simulated using the Weather Research and Forecasting (WRF) Model at 3- and 1-km horizontal grid spacings to investigate the impact of increased resolution on forecasts of convective morphology and its evolution. These simulations were conducted using four microphysics schemes to account for additional sensitivities to the microphysical parameterization. The observed and corresponding simulated systems were manually classified into detailed cellular and linear modes, and the overall morphology depiction and the forecast accuracy of each model configuration were evaluated. In agreement with past studies, WRF was found to underpredict the occurrence of linear modes and overpredict cellular modes at 3-km horizontal grid spacing with all microphysics schemes tested. When grid spacing was reduced to 1 km, the proportion of linear systems increased. However, the increase was insufficient to match observations throughout the evolution of the systems, and the accuracy scores showed no statistically significant improvement. This suggests that the additional linear modes may have occurred in the wrong subtypes, wrong systems, and/or at the wrong times. Accuracy scores were also shown to decrease with forecast length, with the primary decrease in score generally occurring during upscale growth in the early nocturnal period.
Abstract
Nocturnal mesoscale convective systems (MCSs) are important phenomena because of their contributions to warm-season precipitation and association with severe hazards. Past studies have shown that their morphology remains poorly forecast in current convection-allowing models operating at 3–4-km horizontal grid spacing. A total of 10 MCS cases occurring in weakly forced environments were simulated using the Weather Research and Forecasting (WRF) Model at 3- and 1-km horizontal grid spacings to investigate the impact of increased resolution on forecasts of convective morphology and its evolution. These simulations were conducted using four microphysics schemes to account for additional sensitivities to the microphysical parameterization. The observed and corresponding simulated systems were manually classified into detailed cellular and linear modes, and the overall morphology depiction and the forecast accuracy of each model configuration were evaluated. In agreement with past studies, WRF was found to underpredict the occurrence of linear modes and overpredict cellular modes at 3-km horizontal grid spacing with all microphysics schemes tested. When grid spacing was reduced to 1 km, the proportion of linear systems increased. However, the increase was insufficient to match observations throughout the evolution of the systems, and the accuracy scores showed no statistically significant improvement. This suggests that the additional linear modes may have occurred in the wrong subtypes, wrong systems, and/or at the wrong times. Accuracy scores were also shown to decrease with forecast length, with the primary decrease in score generally occurring during upscale growth in the early nocturnal period.
Abstract
A two-dimensional nonhydrostatic version of the NCEP regional Eta Model together with analytic theory are used to examine flow over isolated mountains in numerical simulations using a step-terrain vertical coordinate. Linear theory indicates that a singularity arises in the steady flow over the step corners for hydrostatic waves and that this discontinuity is independent of height. Analytic solutions for both hydrostatic and nonhydrostatic waves reveal a complex behavior that varies with both horizontal and vertical resolution.
Witch of Agnessi experiments are performed with a 2D version of the Eta Model over a range of mountain half-widths. The simulations reveal that for inviscid flow over a mountain using the step-terrain coordinate, flow will not properly descend along the lee slope. Rather, the flow separates downstream of the mountain and creates a zone of artificially weak flow along the lee slope. This behavior arises due to artificial vorticity production at the corner of each step and can be remedied by altering the finite differencing adjacent to the step to minimize spurious vorticity production.
In numerical simulations with the step-terrain coordinate for narrow mountains where nonhydrostatic effects are important, the disturbances that arise at step corners may be of the same horizontal scale as those produced by the overall mountain, and the superposition of these disturbances may reasonably approximate the structure of the continuous mountain wave. For wider mountains, where perturbations are nearly hydrostatic, the disturbances above the step corners have horizontal scales that are much smaller than the overall scale of the mountain and appear as sharp spikes in the flow field. The deviations from the “classic” Witch of Agnesi solution are significant unless the vertical resolution is very small compared to the height of the mountain. In contrast, simulations with the terrain-following vertical coordinate produce accurate solutions provided the vertical grid interval is small compared to the vertical wavelength of the mountain waves (typically at least an order of magnitude larger than the mountain height).
Abstract
A two-dimensional nonhydrostatic version of the NCEP regional Eta Model together with analytic theory are used to examine flow over isolated mountains in numerical simulations using a step-terrain vertical coordinate. Linear theory indicates that a singularity arises in the steady flow over the step corners for hydrostatic waves and that this discontinuity is independent of height. Analytic solutions for both hydrostatic and nonhydrostatic waves reveal a complex behavior that varies with both horizontal and vertical resolution.
Witch of Agnessi experiments are performed with a 2D version of the Eta Model over a range of mountain half-widths. The simulations reveal that for inviscid flow over a mountain using the step-terrain coordinate, flow will not properly descend along the lee slope. Rather, the flow separates downstream of the mountain and creates a zone of artificially weak flow along the lee slope. This behavior arises due to artificial vorticity production at the corner of each step and can be remedied by altering the finite differencing adjacent to the step to minimize spurious vorticity production.
In numerical simulations with the step-terrain coordinate for narrow mountains where nonhydrostatic effects are important, the disturbances that arise at step corners may be of the same horizontal scale as those produced by the overall mountain, and the superposition of these disturbances may reasonably approximate the structure of the continuous mountain wave. For wider mountains, where perturbations are nearly hydrostatic, the disturbances above the step corners have horizontal scales that are much smaller than the overall scale of the mountain and appear as sharp spikes in the flow field. The deviations from the “classic” Witch of Agnesi solution are significant unless the vertical resolution is very small compared to the height of the mountain. In contrast, simulations with the terrain-following vertical coordinate produce accurate solutions provided the vertical grid interval is small compared to the vertical wavelength of the mountain waves (typically at least an order of magnitude larger than the mountain height).
Abstract
A set of mesoscale convective systems (MCSs) was simulated using the Weather Research and Forecasting model with 3-km grid spacing to investigate the skill at predicting convective initiation and upscale evolution into an MCS. Precipitation was verified using equitable threat scores (ETSs), the neighborhood-based fractions skill score (FSS), and the Method of Object-Based Diagnostic Evaluation. An illustrative case study more closely examines the strong influence that smaller-scale forcing features had on convective initiation.
Initiation errors for the 36 cases were in the south-southwest direction on average, with a mean absolute displacement error of 105 km. No systematic temporal error existed, as the errors were approximately normally distributed. Despite earlier findings that quantitative precipitation forecast skill in convection-parameterizing simulations is a function of the strength of large-scale forcing, this relationship was not present in the present study for convective initiation. However, upscale evolution was better predicted for more strongly forced events according to ETSs and FSSs. For the upscale evolution, the relationship between ETSs and object-based ratings was poor. There was also little correspondence between object-based ratings and the skill at convective initiation. The lack of a relationship between the strength of large-scale forcing and model skill at forecasting initiation is likely due to a combination of factors, including the strong role of small-scale features that exert an influence on initiation, and potential errors in the analyses used to represent observations. The limit of predictability of individual convective storms on a 3-km grid must also be considered.
Abstract
A set of mesoscale convective systems (MCSs) was simulated using the Weather Research and Forecasting model with 3-km grid spacing to investigate the skill at predicting convective initiation and upscale evolution into an MCS. Precipitation was verified using equitable threat scores (ETSs), the neighborhood-based fractions skill score (FSS), and the Method of Object-Based Diagnostic Evaluation. An illustrative case study more closely examines the strong influence that smaller-scale forcing features had on convective initiation.
Initiation errors for the 36 cases were in the south-southwest direction on average, with a mean absolute displacement error of 105 km. No systematic temporal error existed, as the errors were approximately normally distributed. Despite earlier findings that quantitative precipitation forecast skill in convection-parameterizing simulations is a function of the strength of large-scale forcing, this relationship was not present in the present study for convective initiation. However, upscale evolution was better predicted for more strongly forced events according to ETSs and FSSs. For the upscale evolution, the relationship between ETSs and object-based ratings was poor. There was also little correspondence between object-based ratings and the skill at convective initiation. The lack of a relationship between the strength of large-scale forcing and model skill at forecasting initiation is likely due to a combination of factors, including the strong role of small-scale features that exert an influence on initiation, and potential errors in the analyses used to represent observations. The limit of predictability of individual convective storms on a 3-km grid must also be considered.
