Abstract

Using data from the 6 July 2015 PECAN case study, this paper provides the first objective assessment of how the assimilation of ground-based remote sensing profilers affects the forecasts of bore-driven convection. To account for the multiscale nature of the phenomenon, data impacts are examined separately with respect to (i) the bore environment, (ii) the explicitly resolved bore, and (iii) the bore-initiated convection. The findings from this work suggest that remote sensing profiling instruments provide considerable advantages over conventional in situ observations, especially when the retrieved data are assimilated at a high temporal frequency. The clearest forecast improvements are seen in terms of the predicted bore environment where the assimilation of kinematic profilers reduces a preexisting bias in the structure of the low-level jet. Data impacts with respect to the other two forecast components are mixed in nature. While the assimilation of thermodynamic retrievals from the Atmospheric Emitted Radiance Interferometer (AERI) results in the best convective forecast, it also creates a positive bias in the height of the convectively generated bore. Conversely, the assimilation of wind profiler data improves the characteristics of the explicitly resolved bore, but tends to further exacerbate the lack of convection in the control forecasts. Various dynamical diagnostics utilized throughout this study provide a physical insight into the data impact results and demonstrate that a successful prediction of bore-driven convection requires an accurate depiction of the internal bore structure as well as the ambient environment ahead of it.

1. Introduction

Convectively generated bores forming in the Great Plains of the United States are an integral part of the nighttime convection that occurs over this region (Crook et al. 1990; Koch and Clark 1999; Parker 2008; Blake et al. 2017; Haghi et al. 2017; Parsons et al. 2019; Grasmick et al. 2018; Haghi et al. 2019). Nevertheless, the representation of bores in numerical weather prediction (NWP) models is very challenging due to the presence of multiple processes driving their initiation and subsequent evolution. For example, as with any convective system, numerical forecasts need to accurately capture both the location and timing of convection initiation (CI) as well as how the simulated convection evolves toward the late afternoon/early evening hours. Assuming that the aforementioned convective activity is simulated well, successfully predicting the formation of a bore further depends on how well the model resolves the interactions between the convectively generated cold pools and the flow within the stable boundary layer (SBL). Capturing the evolution of the bore, on the other hand, is contingent upon the model’s ability to adequately simulate the nocturnal low-level jet (LLJ), whose curvature aids in trapping wave energy behind the leading edge of the bore (e.g., Karyampudi et al. 1995). Finally, the generation of bore-driven convection is sensitive to both the structure of the bore as well as the thermodynamic properties of the ambient environment (Crook 1986).

Given these challenges, high quality initial conditions in and above the SBL are necessary to obtain accurate forecasts of the bores and their influence on nocturnal convection. The low spatiotemporal resolution of current observational networks is insufficient to capture the salient mesoscale features in the lowest 5 km of the atmosphere (Geerts et al. 2017). The solution put forth by several reports from the National Research Council (NRC 2009, 2010, 2012) was to design a network of high-frequency thermodynamic and kinematic profilers capable of conducting continuous unattended measurements of the lower troposphere. Advances in remote sensing have already led to the development of such instruments, with several early studies reporting that profilers provide considerable advantages over conventional radiosondes1 (e.g., Hogg et al. 1983a,b; Bleck and Brummer 1984; Westwater et al. 1984). A notable characteristic of these original profiling systems, which utilized VHF and UHF Doppler radars, was their ability to provide wind measurements with high vertical resolution—an important impetus for the later development of the NOAA Wind Profiler Network (e.g., Weber et al. 1990). Having recognized the potential benefits of the newly introduced kinematic profilers, the meteorological community started to examine the feasibility of these new measurements for different NWP applications. The seminal paper of Kuo et al. (1987), for instance, demonstrated that the assimilation of wind profilers can increase the short-range (0–48 h) skill associated with mesoscale NWP models. Benefits from assimilating wind profilers were also found in several other papers (e.g., Cram et al. 1991; Smith and Benjamin 1993; Guo et al. 2000) and eventually led to the operational assimilation of these kinematic profilers (Bouttier 2001; Benjamin et al. 2004). More recently, several research groups have also considered the assimilation of Doppler wind lidar (DWL) data. The ability of DWLs to provide high-resolution wind information in the lower parts of the BL and in the near-storm environment has led to improvements in the convective forecasts from high-resolution NWP models (Zhang and Pu 2011; Kawabata et al. 2014).

Although early profiling systems, such as the prototype profiler system discussed in Hogg et al. (1983a), did have radiometric profiling capabilities, the vertical resolution of the retrieved temperature and moisture profiles was much coarser than the corresponding wind profiles and comparable to the resolution obtained from satellites (Bleck and Brummer 1984). Later developments in remote sensing technology gave rise to new profiling strategies that allowed for a detailed description of the thermodynamics in the lower troposphere. Examples for some of these instruments include the Raman lidar (Melfi and Whiteman 1985; Melfi et al. 1989) and the Differential Absorption Lidar (DIAL; Browell et al. 1998; Weckwerth et al. 2016), both of which are based on active remote sensing techniques. The NASA Lidar Atmosphere Sensing Experiment (LASE; Moore et al. 1997) is one of the most widely used DIAL instruments and has been deployed throughout multiple field campaigns. The assimilation of LASE data has also been shown to improve the forecasts of hurricanes (Kamineni et al. 2003, 2006) and deep moist convection over the Great Plains (Whiteman et al. 2006). Parallel with the steady progress in lidar technology, the development of passive infrared sensors has also made it possible to obtain simultaneous profiles of temperature and moisture in the lower troposphere. One such instrument is the Atmospheric Emitted Radiance Interferometer (AERI; Revercomb et al. 1988; Feltz et al. 2003; Knuteson et al. 2004a,b; Turner and Löhnert 2014) whose experimental assimilation has improved the analysis and forecasts of large-scale extratropical systems (Hartung et al. 2011; Otkin et al. 2011) and deep moist convection (Coniglio et al. 2019; Degelia et al. 2019).

Despite the encouraging results from the previous studies, further work is still needed to justify a national observational network of profilers that can be assimilated routinely in NWP models. A critical component of establishing this justification is assessing the potential benefit of these novel instruments in predicting high-impact weather, such as nocturnal convection. To address this requirement, the Plains Elevated Convection At Night (PECAN) field campaign, which took place in the summer of 2015, utilized a rich set of thermodynamic and kinematic profilers that targeted nocturnal convective events (Geerts et al. 2017). In an effort to understand whether these state-of-the-art profiling instruments can adequately sample the broad spectrum of processes occurring in the nocturnal environment, this paper examines how their assimilation affects the numerical prediction of the bore-driven convection event from 6 July 2015. The performance of the PECAN profilers is assessed against the operational observation network as well as the high-frequency surface and radiosonde observations collected during this PECAN intensive observing period (IOP). The primary objective of this paper is to understand whether the assimilation of thermodynamic and kinematic data from these PECAN profilers can improve the simulation of the wide range of processes accompanying bore-initiated convection. To address this goal, verification results are presented separately with respect to (i) the environment in which the bore forms, (ii) the explicitly resolved bore and (iii) the bore-initiated nocturnal convection. The benefits from remote sensing profilers are expected to be maximized if their assimilation leads to the simultaneous improvement of all three forecast components.

Studying the impact of remote sensing instruments on the numerical prediction of bores and bore-driven nocturnal convection is a novel aspect of this work, which has been previously hampered due the lack of appropriate verification metrics. The object-based convective outflow algorithm of Chipilski et al. (2018) was an important development that provided the necessary tools for conducting the present investigation. As detailed in section 3b, the aforementioned algorithm has been further extended to accommodate a neighborhood verification of the ensemble bore forecasts as well as diagnostic tools to examine the properties of the simulated bores in the context of the ambient environment. The results reported herein complement the work of Degelia et al. (2019) who demonstrated the ability of PECAN profilers to improve the forecasts of nocturnal CI. Despite their common focus on assimilating novel remote sensing profilers, the two studies look at different aspects of the nocturnal environment and differ in their underlying experimental designs. More specifically, the assimilation of PECAN data in our paper is done in a data addition framework, wherein each new instrument type is added on top of the current conventional observation network. This approach allows us to assess the relative strengths and weaknesses of different instrument types.

The rest of this article is organized as follows: section 2 provides a brief discussion of the case study used in our real-time experiments and highlights the relevancy of the assimilated data with respect to the underlying research objectives. Section 3 describes the experimental design as well as the methodology used for the objective verification of the ensemble forecasts. Data impacts with respect to the three forecast components are discussed separately in sections 4, 5, and 6 and then reconciled in section 7. The paper concludes with a brief discussion regarding the implications of our findings and provides some suggestions on how the representation of nocturnal convection can be improved in future NWP systems.

2. Case study overview and PECAN data availability

The forecast impact of assimilating novel PECAN observations is assessed with respect to the 6 July 2015 case study (IOP20). The radar overview in Fig. 1 shows that there were three regions of afternoon convection [northwestern parts of Nebraska as well as across the Nebraska–South Dakota and South Dakota–Minnesota borders] that grew upscale during the night and eventually merged into a large mesoscale convective system (MCS). During the development of the nocturnal SBL, the strength of the leading convective line across the northern parts of Nebraska decreased and new convective activity developed upstream of the MCS. The presence of such convective initiation is commonly referred to as discrete propagation and has been shown to play an important role in the maintenance of nocturnal convective systems (Crook and Moncrieff 1988; Fovell et al. 2006; Bodine et al. 2017). For this PECAN case, the discrete propagation of the MCS was associated with two separate episodes of bore-initiated convection: one between 0500 and 0615 UTC a second one between 0630 and 0800 UTC (refer to purple ellipses in Fig. 1). The increased longevity of the MCS as a result of the two initiation episodes makes this case especially relevant for the objectives of our study.

Fig. 1.

Radar evolution of the 6 Jul 2015 nocturnal MCS case study based on a 1-km NEXRAD reflectivity mosaic (http://www2.mmm.ucar.edu/imagearchive). The locations corresponding to the first and second episodes of bore-initiated convection are indicated with a magenta ellipse in (d) and (g), while the horizontal scale of the maps is shown in the upper-right corner of (a).

Fig. 1.

Radar evolution of the 6 Jul 2015 nocturnal MCS case study based on a 1-km NEXRAD reflectivity mosaic (http://www2.mmm.ucar.edu/imagearchive). The locations corresponding to the first and second episodes of bore-initiated convection are indicated with a magenta ellipse in (d) and (g), while the horizontal scale of the maps is shown in the upper-right corner of (a).

Selecting IOP20 to conduct our data assimilation experiments was largely driven by the large number of observations collected in this PECAN mission. The nocturnal environment on 6 July was sampled by both fixed and mobile PECAN Integrated Sounding Systems (PISAs) as well as all three research aircrafts deployed during the field campaign (refer to Table 1). It is apparent from Fig. 2 that there were two main regions of data collection. The first one was located in the southeastern parts of South Dakota and featured a mobile array of AERI, DWL, wind profiler, radiosonde, and surface instrumentation. These mobile units were able to observe both sides of the approaching cold pool, providing a unique opportunity to examine the accuracy of the bore environment predictions. However, it should be noted that the cold pool in South Dakota was not observed as well as the ambient environment ahead of it due to the inability of certain profiling instruments (e.g., AERI and DWL) to operate in heavy rain conditions. The other location of intensive data collection was the fixed PISA site in Minden, Nebraska (FP4 hereafter), which hosted AERI, wind profiler, radiosonde, and surface instrumentation. Similar to the mobile array, the FP4 site was able to observe the region upstream of the approaching bore and bore-initiated convection. Additionally, the Compact Raman lidar (CRL) mounted on board the University of Wyoming King Air Research Aircraft (UWKA; Wang et al. 2016) sampled the structure of the observed bore in Nebraska (downward-pointing maroon triangles in Fig. 2b) and generated a valuable dataset for verifying the explicit bore forecasts (discussed in section 5).

Table 1.

PECAN instruments assimilated during the 6 Jul 2015 case study. The variables T, Mυ, U, and V stand for temperature, mixing ratio, and the two components of the horizontal wind. The asterisk (*) in the last row indicates that the NSSL’s Mobile Mesonet (NSSL MM) suite consists of 6 vehicles in total—two NSSL vehicles, two NSSL MGAUS vehicles, one CSU MGAUS vehicle, and the NSSL NOXP scout vehicle.

PECAN instruments assimilated during the 6 Jul 2015 case study. The variables T, Mυ, U, and V stand for temperature, mixing ratio, and the two components of the horizontal wind. The asterisk (*) in the last row indicates that the NSSL’s Mobile Mesonet (NSSL MM) suite consists of 6 vehicles in total—two NSSL vehicles, two NSSL MGAUS vehicles, one CSU MGAUS vehicle, and the NSSL NOXP scout vehicle.
PECAN instruments assimilated during the 6 Jul 2015 case study. The variables T, Mυ, U, and V stand for temperature, mixing ratio, and the two components of the horizontal wind. The asterisk (*) in the last row indicates that the NSSL’s Mobile Mesonet (NSSL MM) suite consists of 6 vehicles in total—two NSSL vehicles, two NSSL MGAUS vehicles, one CSU MGAUS vehicle, and the NSSL NOXP scout vehicle.
Fig. 2.

Spatial distribution of the IOP20 PECAN observations during (a) the final analysis time at 0300 UTC (cf. Fig. 3b) and (b) the first episode of bore-initiated convection at 0500 UTC. Observation locations are plotted over a 10-min period centered around these two times. Overlaid on the two panels are composite reflectivity from the Multi-Radar Multi-Sensor (MRMS) system (color shading) and the ensemble mean 2-m mixing ratio from the BASELINE experiment (solid black contours with values in g kg−1).

Fig. 2.

Spatial distribution of the IOP20 PECAN observations during (a) the final analysis time at 0300 UTC (cf. Fig. 3b) and (b) the first episode of bore-initiated convection at 0500 UTC. Observation locations are plotted over a 10-min period centered around these two times. Overlaid on the two panels are composite reflectivity from the Multi-Radar Multi-Sensor (MRMS) system (color shading) and the ensemble mean 2-m mixing ratio from the BASELINE experiment (solid black contours with values in g kg−1).

3. Methods

a. Experimental design

The data assimilation experiments in this study were conducted with an ensemble data assimilation and prediction system based on version 3.7.1 of the Advanced Research version of the Weather Research and Forecasting (WRF) Model (WRF-ARW; Skamarock et al. 2008) and a Gridpoint Statistical Interpolation (GSI)–based ensemble Kalman filter that was enhanced with convective-scale radar data assimilation capabilities (Johnson et al. 2015; Wang and Wang 2017). The model parameterization schemes utilized in our numerical experiments follow the configuration of Johnson and Wang (2017) and are summarized in Table 2. The timeline in Fig. 3b illustrates the experimental design for the data addition experiments. First, an outer 9-km domain (d01; Fig. 3a) was run with initial and lateral boundary conditions based on 20 members from the Global Ensemble Forecast System (GEFS; Wei et al. 2008) and 20 members from the Short-Range Ensemble Forecast (SREF; Du et al. 2014). In the period between 1500 UTC 5 July and 0000 UTC 6 July 2015, conventional data (surface, radiosonde, ship, buoy, and aircraft flight level) from the North American Mesoscale Forecast System Data Assimilation System (NDAS) were assimilated at a 3-h frequency on the d01 domain. Then, the 9-km ensemble analyses were downscaled to create two additional model nests at 3 km (d02) and 1 km (d03). Between 0000 and 0300 UTC, data assimilation was conducted on the highest resolution 1-km (d03) domain and the conventional NDAS observations were assimilated together with radar data at a temporal frequency of 10 min, following Johnson et al. (2017). Finally, the ensemble analyses obtained at 0300 UTC were used to launch a 40-member ensemble forecast for an additional period of 6 h. The model configuration described so far serves as a control experiment and is referred to as BASELINE hereafter. Additional data impact experiments were also conducted, in which PECAN observations from IOP20 were assimilated on the d03 domain. The name of those experiments alongside with optimally tuned2 localization values for EnKF are summarized in Table 3. Arguably, the vertical localization value in LIDAR_VAD is relatively large, especially in view of the rapidly decreasing accuracy of the DWL instrument over height. However, since additional LIDAR_VAD experiments with different horizontal and vertical localization scales resulted in very little forecast impacts (not shown), the optimal localization choices for this experiment were not explored any further in this study.

Table 2.

WRF-ARW Model physics.

WRF-ARW Model physics.
WRF-ARW Model physics.
Fig. 3.

Experimental design. (a) Model domains used for the numerical experiments in this study. The horizontal grid spacing of domains d01, d02, and d03 is 9, 3, and 1 km, respectively. Each model domain consists of 50 model levels distributed according to the default WRF settings. (b) Timeline of the data assimilation cycling and ensemble free forecasts. Based on initial and lateral boundary conditions from GEFS/SREF, a 40-member ensemble forecast is run between 1200 and 1500 UTC to provide the ensemble background for the first data assimilation cycle on the outer 9-km (d01) domain. Conventional (NDAS) observations are then assimilated every 3 h until 0000 UTC 6 Jul 2015, at which point the data assimilation calculations switch to the highest-resolution (d03) domain for another 3 h. During this period, assimilation frequency is reduced to 10 min and the conventional observations are complemented by radar and PECAN IOP20 data. Finally, the 1-km ensemble analyses at 0300 UTC are used to launch 6-h ensemble forecasts in order to evaluate the PECAN data impacts.

Fig. 3.

Experimental design. (a) Model domains used for the numerical experiments in this study. The horizontal grid spacing of domains d01, d02, and d03 is 9, 3, and 1 km, respectively. Each model domain consists of 50 model levels distributed according to the default WRF settings. (b) Timeline of the data assimilation cycling and ensemble free forecasts. Based on initial and lateral boundary conditions from GEFS/SREF, a 40-member ensemble forecast is run between 1200 and 1500 UTC to provide the ensemble background for the first data assimilation cycle on the outer 9-km (d01) domain. Conventional (NDAS) observations are then assimilated every 3 h until 0000 UTC 6 Jul 2015, at which point the data assimilation calculations switch to the highest-resolution (d03) domain for another 3 h. During this period, assimilation frequency is reduced to 10 min and the conventional observations are complemented by radar and PECAN IOP20 data. Finally, the 1-km ensemble analyses at 0300 UTC are used to launch 6-h ensemble forecasts in order to evaluate the PECAN data impacts.

Table 3.

List of the main numerical experiments conducted as part of the 6 Jul 2015 case study. AIRCRAFT contains both NOAA P-3 flight level data and NASA DC-8 LASE mixing ratio profiles, while WIND_PROF only considers the assimilation of the 915 MHz radio wind profiler at the FP4 site. The horizontal and vertical EnKF localization values in the second and third columns are optimally tuned to maximize the skill of the convective forecasts (cf. section 6). Note that the two localization values reported next to BASELINE refer to the conventional and radar observations, respectively.

List of the main numerical experiments conducted as part of the 6 Jul 2015 case study. AIRCRAFT contains both NOAA P-3 flight level data and NASA DC-8 LASE mixing ratio profiles, while WIND_PROF only considers the assimilation of the 915 MHz radio wind profiler at the FP4 site. The horizontal and vertical EnKF localization values in the second and third columns are optimally tuned to maximize the skill of the convective forecasts (cf. section 6). Note that the two localization values reported next to BASELINE refer to the conventional and radar observations, respectively.
List of the main numerical experiments conducted as part of the 6 Jul 2015 case study. AIRCRAFT contains both NOAA P-3 flight level data and NASA DC-8 LASE mixing ratio profiles, while WIND_PROF only considers the assimilation of the 915 MHz radio wind profiler at the FP4 site. The horizontal and vertical EnKF localization values in the second and third columns are optimally tuned to maximize the skill of the convective forecasts (cf. section 6). Note that the two localization values reported next to BASELINE refer to the conventional and radar observations, respectively.

It is important to mention that all PECAN observations were preprocessed before assimilation (cf. Table 4) by closely following the methodology outlined in Degelia et al. (2019). However, a notable difference with the aforementioned study is that the AERI observation errors were not additionally inflated to account for representativeness errors. The decision to retain the original AERIoe error profiles was motivated by the better forecast skill obtained in this particular case study; nevertheless, future work is still needed to understand the optimal ways of assimilating these novel thermodynamic profilers. Some insights into this question are provided in section 6a where the sensitivity of the AERI convective forecasts is tested for different error and assimilation frequency choices. Last, we also note that the wind profilers from the mobile array in South Dakota were discarded from WIND_PROF and PECAN_ALL experiments due to the presence of large systematic errors3.

Table 4.

Preprocessing and error statistics associated with the PECAN observations from IOP20.

Preprocessing and error statistics associated with the PECAN observations from IOP20.
Preprocessing and error statistics associated with the PECAN observations from IOP20.

b. Verification and diagnostics of the bore forecasts

The impact of assimilating PECAN observations on the quality of the bore forecasts was assessed objectively using the convective outflow algorithm of Chipilski et al. (2018). This algorithm tracks convective outflow boundaries simulated during an ensemble forecast and diagnoses their properties (e.g., bore height) over a user-prescribed reference point R, which is chosen to coincide with the location of the verifying bore observations. Although the aforementioned point-based approach provides a quantitative assessment of forecast accuracy, the verification results will likely not be representative as convective outflow boundary are capable of exhibiting pronounced spatial and temporal variability (Haghi et al. 2017). Moreover, the verification statistics could suffer from sampling errors due to the finite size of the ensemble forecasts.

In an effort to make the verification results more statistically robust, this study extends the original formulation of the algorithm to accommodate an ensemble-based neighborhood verification of the bore forecasts. As illustrated in Fig. 4a, this extension is achieved by generating a reference grid close to a verifying bore observation. The reference grid consists of multiple reference points (four in the example from Fig. 4a) from which the algorithm extracts relevant attributes at the time of object passage. The three reference grids generated as part of this work are shown in Fig. 4c; note that their extent corresponds to a neighborhood radius consisting of 99 reference points, which is the largest number of neighboring points used in this work. The first two reference grids are used to verify the bore environment over the mobile array in South Dakota as well as the explicit bore forecasts near the two UWKA intercepts. The third reference grid is located near the region of bore-initiated convection and is used to understand whether the PECAN impacts are due to changes in the mesoscale ambient environment or the characteristics of the explicitly resolved bore. Analogous to Chipilski et al. (2018), the calculation of object attributes over a particular reference point (Fig. 4b) is achieved by constructing a cross section oriented along the direction of the pressure gradient force. Note, however, that unlike the original formulation of the algorithm, these cross sections are recentered about the location of the leading prefrontal updraft, which allows us to generate composite density current/bore cross sections derived from multiple ensemble members and reference points.

Fig. 4.

Neighborhood verification of the ensemble bore forecasts. (a) Construction of a reference grid in proximity to a verifying bore observation. (b) Four-dimensional (4D) distance minimization at each reference point to select convective outflow boundaries for analysis. Once these objects are determined, the algorithm generates cross sections oriented parallel to the direction of the surface pressure gradient force (PGF; solid magenta vector) to calculate various object attributes, such as the pre and postbore heights (h0 and h1). (c) Location of the 3 reference grids used for the analysis in this study. Additional information overlaid on this panel figure includes the position of the verifying bore observations (e.g., mobile array, UWKA I2 and I6), the location of the KUEX and KOAX radar sites, and the 30 dBZ MRMS reflectivity during the first episode of bore-initiated convection (0546 UTC; thin solid blue contours). The letters N and S refer to the northern and southern clusters of bore-initiated convection, while I2 and I6 refer to the second and sixth bore intercepts from the UWKA aircraft.

Fig. 4.

Neighborhood verification of the ensemble bore forecasts. (a) Construction of a reference grid in proximity to a verifying bore observation. (b) Four-dimensional (4D) distance minimization at each reference point to select convective outflow boundaries for analysis. Once these objects are determined, the algorithm generates cross sections oriented parallel to the direction of the surface pressure gradient force (PGF; solid magenta vector) to calculate various object attributes, such as the pre and postbore heights (h0 and h1). (c) Location of the 3 reference grids used for the analysis in this study. Additional information overlaid on this panel figure includes the position of the verifying bore observations (e.g., mobile array, UWKA I2 and I6), the location of the KUEX and KOAX radar sites, and the 30 dBZ MRMS reflectivity during the first episode of bore-initiated convection (0546 UTC; thin solid blue contours). The letters N and S refer to the northern and southern clusters of bore-initiated convection, while I2 and I6 refer to the second and sixth bore intercepts from the UWKA aircraft.

Verifying the bore environment forecasts is accomplished by utilizing the flow regime diagrams of Rottman and Simpson (1989). These diagrams make theoretical predictions as to whether a density current intruding into an SBL is capable of generating an atmospheric bore. The primary reason for adopting this methodology is in its relative simplicity. In particular, the flow regime diagram describes the characteristics of the bore environment as a single point in the phase space of two nondimensional parameters – the normalized density current depth (D) and the Froude number (F).

4. Bore environment

The correct use of flow regime diagrams requires their application prior to the generation of an upstream bore disturbance. To ensure this criterion is met, we used time series of surface pressure and temperature from both PECAN observations and model forecasts to determine the morphology of the convective outflow boundary that passed over the mobile array in South Dakota. Analysis of these time series revealed a simultaneous rise in pressure and decrease in temperature between 0345 and 0430 UTC (not shown), confirming that the sampled convective outflow boundary had density current characteristics.

a. Observations

For a complete description of the observed flow regime, thermodynamic and kinematic profilers need to simultaneously observe both sides of a density current, which was not always possible during IOP20. To make the best use of the PECAN observations collected on 6 July 2015, data from all relevant mobile sites were combined to estimate the variables needed for the flow regime calculations. Such a framework fits our object-based neighborhood approach and is justified due to the spatial proximity of the mobile instruments.

In the context of the flow regime diagram, the ambient environment ahead of the cold pool is characterized by three variables: the SBL depth (h0), the phase speed of shallow water waves propagating on the SBL inversion (Cgw), and the mean inversion wind speed4 projected in the direction of density current propagation (U0). These flow regime variables were estimates using the available mobile radiosonde, AERI, and DWL observations. It is evident from Fig. 5 that both h0 and U0 increase during the night, coincident with the strengthening SBL inversion and low-level jet (LLJ). It is worth pointing out that the temporal evolution of Cgw follows closely that of h0 and is not displayed here for brevity. Note that observations from different mobile sites are averaged over a 45-min time window prior to the density current arrival (gray boxes in Fig. 5), resulting in a single estimate from each mobile site. A closer examination of the data during this period reveals large regional differences in the observed ambient environment, with h0 varying between 350 and 850 m and U0 exhibiting a bimodal distribution.

Fig. 5.

Observations of the ambient bore environment from the mobile array in southeastern South Dakota showing the temporal evolution of (a) the SBL depth (h0) and (b) the mean inversion wind projected in the direction of density current propagation (U0). The cross markers in the time series show Mobile PISA (MP) estimates derived from either (a) AERI or (b) Doppler wind lidar retrievals. The gray box between 0345 and 0430 UTC highlights the 45-min time period used for calculating the observed flow regime in section 4.

Fig. 5.

Observations of the ambient bore environment from the mobile array in southeastern South Dakota showing the temporal evolution of (a) the SBL depth (h0) and (b) the mean inversion wind projected in the direction of density current propagation (U0). The cross markers in the time series show Mobile PISA (MP) estimates derived from either (a) AERI or (b) Doppler wind lidar retrievals. The gray box between 0345 and 0430 UTC highlights the 45-min time period used for calculating the observed flow regime in section 4.

One of the salient features of IOP20 was the spatial collocation of several instruments from the mobile array in South Dakota, which allowed us to check the quality of the retrieved ambient flow regime variables. As an example, MP1 (CLAMPS) simultaneously operated an AERI, a DWL and a mobile radiosonde; the radiosonde data made it possible to assess the accuracy of the first two remote sensing instruments. The favorable agreement between the two MP1 U0 estimates (green colors in Fig. 5b) does not only demonstrate the DWL’s ability to accurately measure the winds within the SBL, but also provides further evidence for the existence of regional variabilities in the observed LLJ. By contrast, the SBL depth derived from the MP1 AERI retrievals is nearly half of the corresponding radiosonde value. Further comparison of these two instruments showed that the lower AERI h0 value arises due to a stronger vertical gradient in the virtual potential temperature, which acts to lower the SBL depth derived from the Haghi et al. (2015) method. Although the h0 estimate from the MP1 AERI instrument is not too far from the one derived at the nearby MP2 site, we decided to discard it from any subsequent flow regime calculations due to the superior accuracy of the collocated MP1 radiosonde measurements.

The density current properties needed for the flow regime calculations are its depth (d0) and propagation speed (Cg). For consistency with the object-based algorithm of Chipilski et al. (2018), d0 is taken to be the height at which the virtual potential temperature in the density current profile equals its corresponding surface value in the ambient environment. Using the value of d0, the density current’s propagation speed Cg is defined as

 
Cg=gd0ΔρρaαU0,
(1)

where g is the gravitation acceleration, Δρ = ρdρa is the difference in surface air density between the density current and the ambient environment, and α = 0.75 is a correction coefficient that accounts for the slowing of the cold pool as a result of the ambient head wind U0 (Liu and Moncrieff 1996). Estimates of the density current properties were only based on the MG25 (NSSL2) and MG3 (CSU) mobile radiosonde units as heavy precipitation following the passage of the cold pool forced the two AERI instruments to discontinue their data collection process. It is important to remark that while the location of the NSSL2 and CSU units coincided throughout IOP20 (Fig. 4), the balloon launches from these two mobile units occurred at different times. Due to the slowly varying nature of the SBL, the different launching times had very little impact on the ambient flow regime variables (Fig. 5). However, the values in Table 5 indicate that they inevitably caused large deviations in terms of the observed density current properties.

Table 5.

Density current and flow regime characteristics according to the radiosonde observations from the MG2 (NSSL2) and MG3 (CSU) mobile sites. The variables d0 and Cg refer to the depth and propagation speed of the cold pool, while D and F denote the nondimensional density current depth and Froude number. The flow regime classification in the last column follows Rottman and Simpson (1989).

Density current and flow regime characteristics according to the radiosonde observations from the MG2 (NSSL2) and MG3 (CSU) mobile sites. The variables d0 and Cg refer to the depth and propagation speed of the cold pool, while D and F denote the nondimensional density current depth and Froude number. The flow regime classification in the last column follows Rottman and Simpson (1989).
Density current and flow regime characteristics according to the radiosonde observations from the MG2 (NSSL2) and MG3 (CSU) mobile sites. The variables d0 and Cg refer to the depth and propagation speed of the cold pool, while D and F denote the nondimensional density current depth and Froude number. The flow regime classification in the last column follows Rottman and Simpson (1989).

To provide a fair estimate of the bore environment that integrates all available mobile observations in South Dakota, we generated a set of possible flow regimes by randomly combining the values from each of the five flow regime variables. In doing so, special care was taken to account for any theoretical relationships between them. For example, low (high) values of h0 and d0 were only paired with low (high) values of Cgw and Cg. The red crosses in Fig. 6 display the location of all possible flow regimes alongside with the individual flow regime estimates from the CSU and NSSL2 radiosonde sites (refer to the cyan and pink rectangles in Fig. 6 as well as the numbers in Table 5). The observations collected from the mobile array in South Dakota provide strong evidence that the flow regime during IOP was partially blocked, with the majority of the theoretical bore amplitudes (dashed lines) ranging from 2 to 4. According to Rottman and Simpson (1989), these values are associated with a moderately turbulent type B bore. It is interesting to remark that the position of the CSU and NSSL2 estimates in phase space (relative to the set of possible flow regime values) reveals that most of the variability in the observed bore environment comes from the heterogeneous properties of the density current as the two radiosonde sites produce very similar observations of the ambient environment (cf. blue and red dots in Fig. 5).

Fig. 6.

Verification of the bore environment forecasts. The color shading displays the ensemble forecasts from different experiments based on the nearest 49 reference points from the northernmost reference grid in Fig. 4 (number of ensemble members contributing to each experiment, denoted by S, is shown in the upper-left corner of each panel). The flow regime forecasts are verified against individual flow regime estimates from the NSSL2 (pink rectangle) and CSU (cyan rectangle) radiosonde units as well as the set of possible flow regime values derived from all available mobile units in South Dakota (red crosses; see text for more details).

Fig. 6.

Verification of the bore environment forecasts. The color shading displays the ensemble forecasts from different experiments based on the nearest 49 reference points from the northernmost reference grid in Fig. 4 (number of ensemble members contributing to each experiment, denoted by S, is shown in the upper-left corner of each panel). The flow regime forecasts are verified against individual flow regime estimates from the NSSL2 (pink rectangle) and CSU (cyan rectangle) radiosonde units as well as the set of possible flow regime values derived from all available mobile units in South Dakota (red crosses; see text for more details).

b. Forecast impacts

The impact from assimilating PECAN observations on the forecasted bore environment was evaluated for 8 neighborhood radii in order to investigate whether results are consistent across different scales of motion. Since the obtained verification statistics were found to be largely invariant to the chosen neighborhood radius (e.g., Fig. 7a), the discussion herein is restricted to a representative grid containing 49 reference points. Examination of the ensemble forecasts from BASELINE (Fig. 6a) reveals that the control experiment is able to correctly predict a partially blocked flow regime. The presence of two distinct forecast modes, located at theoretical bore amplitudes of ~2 and ~2.7, is also consistent with the tendency of the flow regime observations to cluster around certain portions of the phase space. It should be also noted that while the ensemble spread in BASELINE covers most of the possible flow regime values, the forecast range appears to be slightly displaced toward the origin of the diagram. Improvements in the flow regime predictions are only visible after assimilating observation types that contain wind information [i.e., in the LIDAR_VAD (Fig. 6e), WIND_PROF (Fig. 6f), RADIOSONDE (Fig. 6g), and SURFACE (Fig. 6h) forecasts]. Out of the aforementioned experiments, the best forecast performance is seen in RADIOSONDE and LIDAR_VAD, for which the number of ensemble members below the lower range of possible flow regime values is considerably reduced. In the case of LIDAR_VAD, the ensemble spread is also better aligned with the direction of the observed flow regime uncertainty, which helps increase the ensemble probability over the cluster of possible flow regime values in proximity to the NSSL2 site. As one might expect, the assimilation of all available observations in PECAN_ALL (Fig. 6b) inherits a lot of the positive changes from the well-performing data addition experiments, including the lower likelihood for an undular bore (amplitude less than 2) as well as the larger number of ensemble members over the NSSL2 cluster. Nevertheless, a shortcoming of the PECAN_ALL forecasts is that they are slightly more confident compared to the empirically derived observation uncertainty and produce only a single mode in the forecasted flow regime.

Fig. 7.

Understanding how the assimilated PECAN data affects the forecasted bore environment. (a) Percentage change differences between PECAN_ALL and BASELINE for all 5 flow regime variables and 8 different neighborhood radii using (from left to right) 9, 16, 25, 36, 49, 64, 81, and 99 reference points. (b) A composite density current cross section based on the northernmost reference grid in Fig. 4 with 99 reference points. The buoyancy field from PECAN_ALL is color shaded, while its difference relative to BASELINE is contoured in red (contours plotted every 1 × 10−2 m s−2, starting from ±1 × 10−2 m s−2). The B = −5 × 10−2 m s−2 value from BASELINE is marked with a heavy solid blue line as a reference. Additionally, the solid (dashed) black contours display the virtual potential temperature from PECAN_ALL (BASELINE), which is used for the calculation of the density current and SBL depths (solid black and gray horizontal lines for BASELINE and PECAN_ALL, respectively). (c) Ambient wind speed from BASELINE (solid black line) and PECAN_ALL (dashed black line) projected in the direction opposite to density current propagation and averaged over the last 10 km of the density current cross section in (b). The wind profile from the corresponding radiosonde and Doppler wind lidar data is overlaid using gray color shading. Note that the verifying observations are averaged between 0345 and 0430 UTC, consistent with the gray box in Fig. 5. The width of the gray color shading shows the variability over this 45-min period.

Fig. 7.

Understanding how the assimilated PECAN data affects the forecasted bore environment. (a) Percentage change differences between PECAN_ALL and BASELINE for all 5 flow regime variables and 8 different neighborhood radii using (from left to right) 9, 16, 25, 36, 49, 64, 81, and 99 reference points. (b) A composite density current cross section based on the northernmost reference grid in Fig. 4 with 99 reference points. The buoyancy field from PECAN_ALL is color shaded, while its difference relative to BASELINE is contoured in red (contours plotted every 1 × 10−2 m s−2, starting from ±1 × 10−2 m s−2). The B = −5 × 10−2 m s−2 value from BASELINE is marked with a heavy solid blue line as a reference. Additionally, the solid (dashed) black contours display the virtual potential temperature from PECAN_ALL (BASELINE), which is used for the calculation of the density current and SBL depths (solid black and gray horizontal lines for BASELINE and PECAN_ALL, respectively). (c) Ambient wind speed from BASELINE (solid black line) and PECAN_ALL (dashed black line) projected in the direction opposite to density current propagation and averaged over the last 10 km of the density current cross section in (b). The wind profile from the corresponding radiosonde and Doppler wind lidar data is overlaid using gray color shading. Note that the verifying observations are averaged between 0345 and 0430 UTC, consistent with the gray box in Fig. 5. The width of the gray color shading shows the variability over this 45-min period.

c. Examination of the forecast impacts

Understanding why the assimilation of kinematic profilers contributes to improvements in the bore environment forecasts requires a careful examination of the relative changes in different flow regime variables. The bar plot in Fig. 7 shows a summary of those changes for the PECAN_ALL experiment which produced the largest changes with respect to BASELINE. The most significant discrepancy between the two experiments concerns the magnitude of the inversion wind (U0), which is ≥30% larger in PECAN_ALL for all of the examined neighborhood radii. The composite wind profiles in Fig. 7c confirm this finding and show that the PECAN_ALL forecasts are in much better agreement with the verifying DWL and radiosonde observations. Other notable differences between the two experiments include an 8%–10% reduction in h0 and a corresponding 10%–15% reduction in Cgw. The lower h0 and Cgw values in PECAN_ALL are caused by the smaller static stability in the SBL (cf. the spacing of the first three isentropes in Fig. 7b), which acts to lower the height at which the θυ gradient exceeds its threshold value in the h0 retrieval method (Haghi et al. 2015). On the other hand, the properties of the cold pools in BASELINE and PECAN_ALL are nearly identical, most likely due to (i) the lack of observations within the cold pool and (ii) the negligible moisture differences ahead of the cold pool (which is opposite to our findings in section 6). The small decrease in the strength of the cold pool (d0 and Cg) in PECAN_ALL can be explained by the slightly larger buoyancy values in the upper parts of the cold pool (red contours in Fig. 7b).

Taken as a whole, the relative experimental differences in individual flow regime variables would cause a simultaneous increase in both D and F, shifting the PECAN_ALL’s forecast toward the upper-right parts of the flow regime diagram and alleviating the amplitude bias in the BASELINE experiment (Fig. 6a). The fact that the largest discrepancy between the two experiments pertains to the representation of the LLJ also explains why the assimilation of kinematic information is crucial in improving the bore environment predictions.

5. Explicitly resolved bore

The purpose of the reference grid generated in proximity to the two UWKA intercepts (red dots in Fig. 4c) is to investigate how the assimilated PECAN observations impact the explicit bore forecasts near the FP4 site. In this section, mixing ratio cross sections from the object-based algorithm are compared to UWKA-derived moisture profiles to determine the ability of the forecasts to correctly simulate the observed bore structure during the two verification times. Forecast errors are also quantified objectively for different neighborhood radii using the convective outflow algorithm of Chipilski et al. (2018). The SBL depth in the UWKA retrievals and in the composite model cross sections is defined based on the 11.25 g kg−1 mixing ratio contour, which was objectively found to provide the best description of bore structure in observations and model simulations. The values of h0 and h1 are determined by taking the average height of the 11.25 g kg−1 contour in the subcritical and supercritical portions of the bore. Note that the use of remotely sensed mixing ratio profiles to define the SBL depth was also adopted in Johnson et al. (2018) and Johnson and Wang (2019).

a. UWKA observations and forecast impacts

The height–time resolved mixing ratio (Mυ) profiles in Fig. 8 indicate that the observed bore experiences a rapid evolution over the 50-min time window between the two UWKA intercepts: the postbore height (h1) increases from 1.3 to nearly 2 km AGL and is coincident with an increase in the prebore SBL height (h0) from 450 to 600 m. As a result, the amount of bore lifting (Δh = h1h0) increases significantly between UWKA I2 and UWKA I6 and likely explains the onset of bore-initiated convection around the second verification time.

Fig. 8.

Evolution of the observed bore in Nebraska based on moisture profiles from UWKA’s Compact Raman lidar (CRL) system. The heavy black lines denote the 11.25 g kg−1 mixing ratio contour used to approximate the height of the SBL in the supercritical and subcritical portions of the bore (see text for more details).

Fig. 8.

Evolution of the observed bore in Nebraska based on moisture profiles from UWKA’s Compact Raman lidar (CRL) system. The heavy black lines denote the 11.25 g kg−1 mixing ratio contour used to approximate the height of the SBL in the supercritical and subcritical portions of the bore (see text for more details).

The composite cross sections in Fig. 9 are created in proximity to UWKA I2 and examine the structure of the simulated bore during its early evolution. First, it is clear that the control experiment (Fig. 9a) tends to underestimate (overestimate) the amount of moisture in the prebore (postbore) environment. More pronounced improvements in the forecasted bore environment are visible only after assimilating all available PECAN observations (Fig. 9b), which causes an increase in the prebore SBL mixing ratio by ~0.6 g kg−1. The enhanced SBL moisture in PECAN_ALL also corrects a slight negative bias in the prebore height (h0; compare the thick blue contour and the rightmost magenta line in Fig. 9a). As far as the structure of the simulated bore is concerned, the PECAN impacts are most evident with respect to the numerically predicted bore height (h1), which is overestimated in BASELINE by some 300–400 m (cf. the heavy blue contour and leftmost magenta line in Fig. 9a). Most of the data assimilation experiments exacerbate this error, with AERI and PECAN_ALL featuring the largest deviations from the UWKA-derived h1 estimate. It is perhaps interesting to note that the postbore SBL increases in AIRCRAFT and SURFACE are most prominent downstream of the hydraulic jump, in contrast to AERI and PECAN_ALL where the largest changes occur immediately downstream of the bore’s leading edge. On the other hand, WIND_PROF (Fig. 9f) shows a strong deviation from the other data assimilation experiments—it decreases h1 by ~200–400 m and, consequently, nearly halves the bore height bias in BASELINE. Similar to AERI and PECAN_ALL, the largest changes in the WIND_PROF-predicted SBL height occur very close to the hydraulic jump. Insofar as the other experiments are concerned, LIDAR_VAD and RADIOSONDE produce the smallest differences in the explicitly resolved bore relative to BASELINE. Whereas the absence of DWL profiles upstream of the bore justifies the lack of visible forecast impacts in LIDAR_VAD, a possible explanation for the small RADIOSONDE-BASELINE differences is the coarse temporal resolution of the FP4 radiosonde measurements compared to the other observation types.

Fig. 9.

Verification of the explicit bore forecasts near UWKA I2 (~0500 UTC). (a) Mixing ratio (color shading) and virtual potential temperature (solid black contours) from the BASELINE experiment. The heavy solid blue contour marks the SBL depth, while the two horizontal lines correspond to the UWKA-derived prebore and postbore SBL heights. Note that all forecasted SBL heights are calculated using the 11.25 g kg−1 mixing ratio value, consistent with Fig. 8. (b)–(h) Mixing ratio differences associated with various data assimilation experiments. The virtual potential temperature and SBL height from BASELINE are overlaid as thin and heavy dashed black contours, respectively.

Fig. 9.

Verification of the explicit bore forecasts near UWKA I2 (~0500 UTC). (a) Mixing ratio (color shading) and virtual potential temperature (solid black contours) from the BASELINE experiment. The heavy solid blue contour marks the SBL depth, while the two horizontal lines correspond to the UWKA-derived prebore and postbore SBL heights. Note that all forecasted SBL heights are calculated using the 11.25 g kg−1 mixing ratio value, consistent with Fig. 8. (b)–(h) Mixing ratio differences associated with various data assimilation experiments. The virtual potential temperature and SBL height from BASELINE are overlaid as thin and heavy dashed black contours, respectively.

During the second verification time (UWKA I6), the ambient moisture environment in BASELINE (Fig. 10a) is represented much better, which in turn also improves the structure of the explicitly resolved bore. Furthermore, the experimental differences at UWKA I6 are considerably smaller than UWKA I2; only the AERI (Fig. 10c), WIND_PROF (Fig. 10f), and PECAN_ALL (Fig. 10b) experiments display visible deviations from BASELINE. Despite the apparent reduction in errors, however, none of the ensemble forecasts reproduces the rapid bore evolution depicted in UWKA’s moisture profiles. Given that the experimental differences between the two verification times remain similar, the slower increase of the numerically simulated bore height acts to change the sign of the PECAN impacts between UWKA I2 and I6. For instance, the strongly negative impact in AERI at UWKA I2 (Fig. 9c) on h1 transforms into a marginal improvement at UWKA I6 (Fig. 10c). Likewise, the addition of boundary layer moisture in PECAN_ALL is beneficial early in the forecast (Fig. 9b), but it leads to an overestimate of h0 in vicinity of UWKA I6 (Fig. 10b).

Fig. 10.

Verification of the explicit bore forecasts near UWKA I6 (~0550 UTC). The meaning of all symbols remains the same as in Fig. 9.

Fig. 10.

Verification of the explicit bore forecasts near UWKA I6 (~0550 UTC). The meaning of all symbols remains the same as in Fig. 9.

A summary of the average PECAN impacts on the structure of the explicitly resolved bore is presented in Fig. 11 through the mean average errors (MAEs).6 Note that the verification statistics in Fig. 11 are dominated by the forecast performance at UWKA I2 where the error magnitudes and experimental differences are significantly larger relative to UWKA I6. Despite differences in methodology, however, the MAE results in Fig. 11 are largely consistent with the subjective conclusions presented so far. For instance, Fig. 11a confirms the presence of a small positive bias in PECAN_ALL’s prediction of h0, while Fig. 11b highlights that the assimilation of wind profiler (AERI) data has a positive (negative) impact on the predictions of the numerically simulated bore height.

Fig. 11.

Summary of the verification statistics associated with the explicit bore forecasts over UWKA I2 and UWKA I6. Solid lines with filled markers display the mean average errors (MAEs) with respect to the (a) prebore height (h0) and (b) postbore height (h1) for a different number of reference points (x axis). The gray box at the bottom of the two panel figures shows the percentage of statistically different reference points (SDRP) relative to the BASELINE experiment.

Fig. 11.

Summary of the verification statistics associated with the explicit bore forecasts over UWKA I2 and UWKA I6. Solid lines with filled markers display the mean average errors (MAEs) with respect to the (a) prebore height (h0) and (b) postbore height (h1) for a different number of reference points (x axis). The gray box at the bottom of the two panel figures shows the percentage of statistically different reference points (SDRP) relative to the BASELINE experiment.

The structure of a convectively generated bore is important as it affects errors associated with derived variables, such as bore amplitude (S = h1/h0) and bore lifting (Δh = h1h0). Both of these variables play a crucial role in determining the propagation of a bore (e.g., White and Helfrich 2012) as well as its ability to initiate convection (Parsons et al. 2019). Given the additive nature of Δh errors and the fact that changes in h1 are much greater than changes in h0 (e.g., Fig. 9), the accuracy of the bore lifting predictions is mostly determined by how the assimilated PECAN observations affect the height of the explicitly resolved bore. Indeed, Table 6 confirms that the overprediction of h1 in all experiments translates to a positive bias in the Δh forecasts. Analogous to the MAE h1 results (Fig. 11b), the WIND_PROF (AERI) experiments produce the largest positive (negative) impact on the forecasts of bore lifting.

Table 6.

Example PECAN impacts on the forecasted bore structure for a reference grid with 10 points and based on the MAE results in Fig. 11. The numbers in the first row refer to the UWKA-derived bore estimates, while the remaining rows display forecast results from the main numerical experiments presented in this study. The bracketed numbers next to each experiment denote changes relative to BASELINE.

Example PECAN impacts on the forecasted bore structure for a reference grid with 10 points and based on the MAE results in Fig. 11. The numbers in the first row refer to the UWKA-derived bore estimates, while the remaining rows display forecast results from the main numerical experiments presented in this study. The bracketed numbers next to each experiment denote changes relative to BASELINE.
Example PECAN impacts on the forecasted bore structure for a reference grid with 10 points and based on the MAE results in Fig. 11. The numbers in the first row refer to the UWKA-derived bore estimates, while the remaining rows display forecast results from the main numerical experiments presented in this study. The bracketed numbers next to each experiment denote changes relative to BASELINE.

The values summarized in Table 6 also indicate that the bore amplitude predicted in BASELINE (3.26) is very close to the verifying UWKA retrievals (3.10) despite the presence of large biases in the predicted bore height. According to Rottman and Simpson (1989), both of these S values would result in an identical flow regime characterized by the generation of an upstream bore disturbance with moderate amounts of turbulence. The small S error in BASELINE experiment can be explained by noting that (i) S is a ratio of h1 and h0 and (ii) errors in S are relative in nature. Since h0h1, the value of S is expected to be highly sensitive to small changes in the depth of the SBL ahead of the bore. Because the BASELINE experiment produces fairly accurate predictions of h0 at both UWKA intercepts, the large overestimate of h1 only contributes to a small positive bias in S. Similar arguments can be applied to justify the bore amplitude results in the other data assimilation experiments. For example, while the reduction of h1 in WIND_PROF is almost as high as the increase of h1 in AERI (cf. Table 6), the attendant decrease of h0 by 50 m prevents WIND_PROF from correcting the amplitude bias in BASELINE. Likewise, the increase of h0 and h1 in PECAN_ALL is on the same order of magnitude (~100 m), but the higher sensitivity of S to h0 reduces the predicted bore amplitude.

b. Examination of the forecast impacts

The nonlinear dynamics of atmospheric bores makes the interpretation of the PECAN data impact results inherently challenging. This is because their formation and evolution are very sensitive to the properties of the parent convection and its cold pools. Despite these difficulties, though, multiple studies have demonstrated the success of hydraulic theory in explaining the complex behavior of observed atmospheric bores (Koch et al. 1991, 2008a,b). As a result, the analysis here makes use of hydraulic theory to identify where differences in the AERI and WIND_PROF bore forecasts originate from. We focus on these two experiments not only due to their considerable deviations from the BASELINE experiment, but also to emphasize the dependence of the PECAN data impacts on the assimilated instrument types (i.e., thermodynamic versus kinematic).

The characteristics of the bore environments in BASELINE, AERI, and WIND_PROF are investigated by generating cross sections near the verifying UWKA intercepts. It is apparent from the ensemble mean fields in Fig. 12 that the largest differences in the three experiments are due to the depth of the bore-generating density current (d0). The near-surface isentropes between x = −60 and x = −20 km indicate that the cold pool in AERI (WIND_PROF) is stronger (weaker) than BASELINE. According to well-established theoretical results (e.g., Benjamin 1968), changes in the depth of the cold pool (d0) are also correlated to changes in its propagation speed (Cg), which was confirmed by a subjective analysis of several ensemble mean fields. Apart from changes in the cold pool characteristics, the magnitude of the inversion wind (U0) in WIND_PROF is 1–2 m s−1 smaller than BASELINE (Fig. 12b), causing a further decrease in the forecasted Froude number F. Analogous to section 4, the reduction of the mean inversion wind (U0) in WIND_PROF is connected to the ability of the assimilated FP4 wind profiler to correct a bias in the representation of the LLJ (not shown).

Fig. 12.

Flow regime diagnostics on the bore-generating cold pool in Nebraska for the (a) AERI and (b) WIND_PROF experiments. The color shading on the two cross sections shows the ensemble mean inversion wind projected in the direction of density current propagation (U0; color shading), with negative (positive) values indicating flow oriented toward (away from) the density current. Negative (positive) differences in U0 with respect to the BASELINE experiment are displayed as solid blue (red) contours. The virtual potential temperature from AERI/WIND_PROF and BASELINE is plotted with solid and dashed black contours, respectively.

Fig. 12.

Flow regime diagnostics on the bore-generating cold pool in Nebraska for the (a) AERI and (b) WIND_PROF experiments. The color shading on the two cross sections shows the ensemble mean inversion wind projected in the direction of density current propagation (U0; color shading), with negative (positive) values indicating flow oriented toward (away from) the density current. Negative (positive) differences in U0 with respect to the BASELINE experiment are displayed as solid blue (red) contours. The virtual potential temperature from AERI/WIND_PROF and BASELINE is plotted with solid and dashed black contours, respectively.

The aforementioned differences in the bore-generating density current are related to perturbations in the near-surface moisture field at the FP4 site, which are subsequently transported toward this cold pool during the free forecast period (Fig. 13). As a result of their interaction with the cold pool’s updraft, the positive (negative) moisture perturbations in AERI (WIND_PROF) strengthen (weaken) the parent convection and, through changes in precipitation loading, produce a stronger (weaker) cold pool. It is important to remark that while the positive moisture increments in AERI are introduced primarily due to a moisture bias at the beginning of the 1-km DA cycling period, the drying in WIND_PROF arises due to the inherent mass-wind correlations in the ensemble Kalman filter analyses. For instance, the wind perturbations at 0100 UTC are oriented toward the northeast (magenta arrows in Fig. 13l) and act to strengthen the magnitude of the LLJ. Further note that these perturbations are directed perpendicular to the preexisting moisture boundary in central Nebraska (black contours in Fig. 2). Owing to the stronger air advection from the southwest, the SBL in WIND_PROF becomes drier, generating negative mixing ratio perturbations that are subsequently transported toward the bore-generating cold pool in Nebraska (Fig. 13r).

Fig. 13.

Moisture impacts on the explicit bore forecasts for (a)–(j) the AERI and (k)–(t) the WIND_PROF experiments. Color shading refers to ensemble mean differences in mixing ratio relative to BASELINE. Contours are plotted every 2 g kg−1 with solid lime and bisque colors (first contour appears at ±1 g kg−1). Black and magenta arrows show the ensemble mean wind in AERI/WIND_PROF and their difference with BASELINE. The 30 dBZ composite reflectivity and the regions where vertical velocity at 1 km AGL exceeds 0.5 m s−1 are visualized as solid black contours and purple dots. The three red stars on the bottom portion of the panel figures show the location of the FP4 site and the two UWKA intercepts [labeled in (a)], while the heavy solid black line—the position of the density current cross section from Fig. 12. Note that both the mixing ratio and vertical velocity fields are smoothed using a Gaussian filter with σ = 2.0 km, while the moisture and wind fields in the first 7 (last 2) rows are plotted at 100 m (1.5 km) AGL. Furthermore, the bold letters A and F next to the left panels indicate whether the data are taken from the analysis or forecast ensemble.

Fig. 13.

Moisture impacts on the explicit bore forecasts for (a)–(j) the AERI and (k)–(t) the WIND_PROF experiments. Color shading refers to ensemble mean differences in mixing ratio relative to BASELINE. Contours are plotted every 2 g kg−1 with solid lime and bisque colors (first contour appears at ±1 g kg−1). Black and magenta arrows show the ensemble mean wind in AERI/WIND_PROF and their difference with BASELINE. The 30 dBZ composite reflectivity and the regions where vertical velocity at 1 km AGL exceeds 0.5 m s−1 are visualized as solid black contours and purple dots. The three red stars on the bottom portion of the panel figures show the location of the FP4 site and the two UWKA intercepts [labeled in (a)], while the heavy solid black line—the position of the density current cross section from Fig. 12. Note that both the mixing ratio and vertical velocity fields are smoothed using a Gaussian filter with σ = 2.0 km, while the moisture and wind fields in the first 7 (last 2) rows are plotted at 100 m (1.5 km) AGL. Furthermore, the bold letters A and F next to the left panels indicate whether the data are taken from the analysis or forecast ensemble.

6. Bore-initiated convection

The development of bore-initiated convection can be viewed as a manifestation of all multiscale processes driving the evolution of the bore and its surrounding ambient environment. Therefore, the purpose of this section is to perform an objective assessment of the convective forecasts and then link them to the previously discussed PECAN data impacts. Note that our analysis only focuses on the first episode of bore-initiated convection (0500–0615 UTC; Fig. 1d) when the experimental differences were most pronounced.

a. Radar observations and forecast impacts

The convectively generated bore in Nebraska was captured very well in the low-level reflectivity data from the KOAX radar site (Fig. 14), where it shows up as a series of three radar fine lines propagating toward east-southeast. Convection initiation along the bore was discrete and characterized by the occurrence of two main convective clusters. The northern one, denoted as N, started its development around 0500 UTC at the leading edge of the nocturnal MCS. As the bore propagated away from the MCS, convection initiation associated with cluster N shifted ahead of the main convective system. The southern cluster (S) developed approximately 30 min after N and moved in the direction of bore propagation.

Fig. 14.

Equivalent radar reflectivity (lowest elevation angle) from the KOAX radar site (Omaha, NE) depicting the evolution of the first episode of bore-initiated convection for the 6 Jul 2015 case study. The letters N and S show the location of the northern and southern convective clusters. Radar images are generated using the Python Atmospheric Radiation Measurement (ARM) Radar Toolkit (Py-ART; Helmus and Collis 2016).

Fig. 14.

Equivalent radar reflectivity (lowest elevation angle) from the KOAX radar site (Omaha, NE) depicting the evolution of the first episode of bore-initiated convection for the 6 Jul 2015 case study. The letters N and S show the location of the northern and southern convective clusters. Radar images are generated using the Python Atmospheric Radiation Measurement (ARM) Radar Toolkit (Py-ART; Helmus and Collis 2016).

The skill of the convective forecasts is assessed objectively using neighborhood ensemble probability (NEP; Schwartz et al. 2010) with a threshold of 30 dBZ and a neighborhood radius of 30 km. Comparison of the NEP field from BASELINE and the observed 30 dBZ value derived from Multi-Radar Multi-Sensor (MRMS; Zhang et al. 2016) data in Figs. 15a–c shows that the first episode of bore-initiated convection is underpredicted in the control experiment. Despite an initial increase of NEP to 50% near the location of cluster N (Fig. 15a), the BASELINE forecasts do not maintain the newly developed bore convection for a sufficiently long time.

Fig. 15.

Impact of assimilating PECAN observations on the forecasts of bore-initiated convection. The first row shows results for the BASELINE experiment through the neighborhood ensemble probability (NEP; color shading) of model-simulated reflectivity exceeding 30 dBZ with a neighborhood radius of 8 km. The rest of the panel figures display differences in NEP with respect to the other data assimilation experiments. The observed 30 dBZ composite reflectivity from the Multi-Radar Multi-Sensor (MRMS) system is contoured in black.

Fig. 15.

Impact of assimilating PECAN observations on the forecasts of bore-initiated convection. The first row shows results for the BASELINE experiment through the neighborhood ensemble probability (NEP; color shading) of model-simulated reflectivity exceeding 30 dBZ with a neighborhood radius of 8 km. The rest of the panel figures display differences in NEP with respect to the other data assimilation experiments. The observed 30 dBZ composite reflectivity from the Multi-Radar Multi-Sensor (MRMS) system is contoured in black.

The convective impacts from the other data addition experiments are shown as differences in NEP relative to the BASELINE experiment. Examination of these difference fields reveals that LIDAR_VAD (Figs. 15m–o), AIRCRAFT (Figs. 15j–l), and SURFACE (Figs. 15v–x) exert a negligible impact on the convective forecasts. The latter is not so surprising due to (i) the absence of DWL and aircraft data upstream of the forecasted bore convection and (ii) the relatively small impact that surface observations have had on the forecast results so far (cf. sections 4 and 5). The most successful prediction of the observed bore-initiated convection is produced by the AERI experiment (Figs. 15g–i), which is able to (i) increase the NEP values by more than 50% and (ii) successfully capture the initiation and propagation of the southern convective cluster (S). By contrast, assimilation of the FP4 wind profiler data leads to a slight degradation in the convective forecast skill, evident by the reduction in NEP in Figs. 15p–r.

The significant improvement of the convective forecast skill in AERI prompted us to conduct two additional experiments aimed at exploring the sensitivity of the results to alternative DA configurations. In the first one, the AERI profiles were assimilated with static GSI radiosonde error statistics (AERI_SOUNDERR; Fig. 16b). Since radiosonde errors are notably larger than their AERIoe counterparts (refer to Degelia et al. (2019)), AERI_SOUNDERR can be thought of as an experiment which crudely accounts for the errors of representation inherent in the AERIoe profiles. According to Figs. 16a and 16b, both AERI experiments result in similar convective forecasts, although AERI_SOUNDERR tends to slightly overestimate the convective probabilities around cluster S. Nonetheless, additional diagnostics revealed that AERI_SOUNDERR degrades the quality of the bore-related forecasts (not shown), which justifies our choice of the dynamically retrieved error profiles in the control AERI experiment. The goal behind the second sensitivity experiment (AERI_REDUCED; Fig. 16c) was to understand the extent to which the quality of the convective forecasts is affected by the AERI’s assimilation frequency. More specifically, the thermodynamic retrievals in AERI_REDUCED were assimilated at the same frequency as the FP4 radiosonde observations (i.e., every 3 h). The striking resemblance of the AERI_REDUCED and RADIOSONDE (Fig. 16d) forecasts implies that the assimilation of novel AERI data adds no value to the quality of the NWP forecasts unless it is assimilated at a high temporal frequency. Observation space statistics from the FP4 site (not shown) further confirm that the continuous assimilation of AERI profiles over a 3-h period results in a model state that agrees favorably with the collocated radiosonde observations at 0300 UTC.

Fig. 16.

Sensitivity of the AERI convective forecasts to different data assimilation configurations. NEP fields from 0530 UTC are shown for the original (a) AERI and (d) RADIOSONDE experiments as well as the additionally conducted (b) AERI_SOUNDERR and (c) AERI_REDUCED experiments, which test the forecast sensitivity to different observation errors and assimilation frequencies. The rest of the symbols remain the same as in the first row of Fig. 15.

Fig. 16.

Sensitivity of the AERI convective forecasts to different data assimilation configurations. NEP fields from 0530 UTC are shown for the original (a) AERI and (d) RADIOSONDE experiments as well as the additionally conducted (b) AERI_SOUNDERR and (c) AERI_REDUCED experiments, which test the forecast sensitivity to different observation errors and assimilation frequencies. The rest of the symbols remain the same as in the first row of Fig. 15.

Last, we note that the simultaneous assimilation of all PECAN observations yields mixed impacts on the convective forecasts. The higher NEP values to the east of cluster N (Figs. 15d–f) appear to be associated with spurious convection that develops earlier in the PECAN_ALL simulation, which ultimately inhibits the subsequent development of bore-initiated convection over the chosen verification region. Analogous to the results from section 5, the presence of mixed impacts in AERI and WIND_PROF prevents PECAN_ALL from providing the best forecast results.

b. Examination of the forecast impacts

The accuracy of bore-initiated convection depends on the representation of the explicitly resolved bore as well as the ambient environment ahead of it. Verification results presented so far indicate that the assimilation of PECAN profilers affects both of these factors. By generating composite cross sections near the region of bore-initiated convection (aqua dots in Fig. 4c), this section seeks to establish a link between the skill of the convective forecasts and the PECAN impacts discussed so far. Similar to section 5b, the analysis herein focuses on the AERI and WIND_PROF experiments since they produce the largest deviations from the control forecasts.

It is clear from Figs. 17a, 17d, and 17g that the explicitly resolved bores simulated in the three experiments differ from each other. Consistent with the verification results in section 5, the postbore height (h1) is higher in AERI, and lower in WIND_PROF, than in BASELINE. Changes in the ambient moisture environment also align with the verification results from the previous section. For example, the AERI experiment leads to a significant increase in mixing ratio (Mυ) that reaches nearly 2 g kg−1 in the 3–3.5 km AGL layer (Fig. 17d). As a result, the elevated CAPE in this experiment increases by 100–500 J kg−1 (Fig. 17e). Combined with a more pronounced bore lifting and nearly neutral changes in the elevated CIN values (Fig. 17f), conditions in AERI become favorable for the development of stronger and longer-lived bore convection. By contrast, the slight decrease of moisture in WIND_PROF (~0.2–0.3 g kg−1 at 3–3.5 km; Fig. 17g) tends to lower CAPE by ~100 J kg−1 (Fig. 17h) and, together with a slight reduction in the amount of bore lifting (Fig. 17g), results in a lower probability of bore-initiated convection (Figs. 15p–r).

Fig. 17.

Composite cross sections for the BASELINE, AERI, and WIND_PROF experiments generated in the neighborhood of the southern convective cluster (S; refer to the aqua reference points in Fig. 4). (a),(d),(g) The distribution of moisture from these three experiments (mixing ratio for BASELINE and differences in mixing ratio for AERI and WIND_PROF). The SBL height (based on the 11.25 g kg−1 mixing ratio value) is plotted either with a heavy solid blue line for individual experiments or with a heavy dashed black line for the BASELINE experiment in (d) and (g). The prebore and postbore heights from UWKA I6 are overlaid as heavy horizontal lines. The color shading in the second and third columns shows the composite (b),(e),(h) CAPE and (c),(f),(i) CIN fields for the three experiments. Virtual potential temperature is plotted in all panels with solid black contours.

Fig. 17.

Composite cross sections for the BASELINE, AERI, and WIND_PROF experiments generated in the neighborhood of the southern convective cluster (S; refer to the aqua reference points in Fig. 4). (a),(d),(g) The distribution of moisture from these three experiments (mixing ratio for BASELINE and differences in mixing ratio for AERI and WIND_PROF). The SBL height (based on the 11.25 g kg−1 mixing ratio value) is plotted either with a heavy solid blue line for individual experiments or with a heavy dashed black line for the BASELINE experiment in (d) and (g). The prebore and postbore heights from UWKA I6 are overlaid as heavy horizontal lines. The color shading in the second and third columns shows the composite (b),(e),(h) CAPE and (c),(f),(i) CIN fields for the three experiments. Virtual potential temperature is plotted in all panels with solid black contours.

The analysis presented so far serves as a good example of how the assimilated profiling instruments lead to multiscale forecast impacts. The moisture perturbations in AERI and WIND_PROF affect the quality of the convective forecasts by simultaneously modifying the elevated instability in the ambient environment and changing the height of the explicitly resolved bore.

7. Conclusions and discussion

Recent studies have suggested that bores are an inherent component of the nocturnal environment over the Great Plains (Geerts et al. 2017; Haghi et al. 2017), capable of maintaining ongoing nocturnal convection by destabilizing broad regions in their wake (Parsons et al. 2019), and initiating deep convection on their own (Parker 2008; Grasmick et al. 2018). Despite the growing need to adequately represent bores in NWP models, capturing their initiation and evolution depends on the ability of these models to accurately simulate a wide range of multiscale processes, such as the formation of cold pools from late afternoon/early evening convection as well as the simultaneous development of a stable boundary layer and a low-level jet in the nocturnal environment. Since the individual modeling of these processes is already a difficult task on its own (e.g., Weisman et al. 2008), prediction of bore-driven nocturnal convection poses significant challenges to current convective-scale NWP models. One of the possible ways to reduce errors associated with these processes is by improving the model’s initial conditions in a carefully designed multiscale data assimilation system. The present article represents the first attempt to address this idea by assimilating novel kinematic and thermodynamic profilers deployed during the 2015 PECAN campaign. Using data from the 6 July 2015 PECAN case study, we examined whether the initial conditions obtained after assimilating these PECAN profilers can improve various aspects of the bore-driven nocturnal convection, including (i) the environment in which the bore develops, (ii) the characteristics of the explicitly resolved bore, and (iii) the accuracy of the convective forecasts.

For the reader’s convenience, Table 7 presents a summary of all forecast impacts discussed throughout the paper. Evidently, the clearest impacts from assimilating PECAN data refer to the numerically simulated bore environment where nearly all observation types improve the forecasted flow regime. Section 4c demonstrates that these improvements arise from a better representation of the low-level jet that interacts with the bore-generating cold pool.

Table 7.

Summary of the key PECAN impacts for the 6 Jul 2015 case study. The upward- and downward-pointing arrows correspond to positive and negative forecast impacts with respect to BASELINE, while the neutral forecast impacts are denoted with a dash. The number of arrows in each of the three forecast categories provides a subjective ordering of the data assimilation experiments according to the magnitude of their impacts.

Summary of the key PECAN impacts for the 6 Jul 2015 case study. The upward- and downward-pointing arrows correspond to positive and negative forecast impacts with respect to BASELINE, while the neutral forecast impacts are denoted with a dash. The number of arrows in each of the three forecast categories provides a subjective ordering of the data assimilation experiments according to the magnitude of their impacts.
Summary of the key PECAN impacts for the 6 Jul 2015 case study. The upward- and downward-pointing arrows correspond to positive and negative forecast impacts with respect to BASELINE, while the neutral forecast impacts are denoted with a dash. The number of arrows in each of the three forecast categories provides a subjective ordering of the data assimilation experiments according to the magnitude of their impacts.

By contrast, the PECAN impacts on the explicitly resolved bore and bore-initiated convection appear to be mixed and largely dependent on the chosen verification time. For example, while the AERI experiment produces a large positive bias in the predicted bore height early in the forecasts, it also provides the closest match to the verifying UWKA retrievals around the time of convection initiation. Combined with a better depiction of the moisture content in the ambient environment, the AERI experiment results in the best forecast of the observed bore-initiated convection. These mixed forecast impacts can be explained by the fact that bores and bore-initiated convection are associated with motions that act on very small spatiotemporal scales. The parameterization of these subgrid-scale processes inevitably leads to model errors and may even mask any benefits brought by the assimilation of high-frequency PECAN profilers.

Although the assimilation of thermodynamic and kinematic remote sensors appears to be beneficial for the prediction of nocturnal convection in NWP models, there are still a lot of open research questions. One of the them concerns the optimal design of a profiler network. Even though our findings highlight the advantages of assimilating high-frequency profiler data, more work still needs to be done in order to determine the optimal spacing between profiling instruments. In addition, the relative merits of assimilating collocated thermodynamic and kinematic profilers need to be better understood. Past observing system simulation experiments (OSSEs) have reported benefits from such an observation design (e.g., Hartung et al. 2011), but the results from this study suggest that the simultaneous assimilation of kinematic and thermodynamic profilers often leads to mixed forecast results. Therefore, one of the goals of future research should be to determine effective ways of combining the information from collocated profiling instruments. A systematic data impact study featuring 10 bore IOPs from the PECAN field campaign is currently underway to provide more insight into this question and generalize the findings of this paper.

Acknowledgments

This work is primarily supported by NSF Award AGS-1359703. Computing resources were provided by the Yellowstone (ark:/85065/d7wd3xhc) and Cheyenne (doi:10.5065/D6RX99HX) machines at NCAR’s Computational and Information Systems Laboratory, sponsored by the NSF. The authors of this paper would like to thank Samuel Degelia and Aaron Johnson for their technical support as well as the three anonymous reviewers who provided constructive feedback on the early version of this article.

REFERENCES

REFERENCES
Benjamin
,
S. G.
,
B. E.
Schwartz
,
E. J.
Szoke
, and
S. E.
Koch
,
2004
:
The value of wind profiler data in U.S. weather forecasting
.
Bull. Amer. Meteor. Soc.
,
85
,
1871
1886
, https://doi.org/10.1175/BAMS-85-12-1871.
Benjamin
,
T. B.
,
1968
:
Gravity currents and related phenomena
.
J. Fluid Mech.
,
31
,
209
248
, https://doi.org/10.1017/S0022112068000133.
Blake
,
B. T.
,
D. B.
Parsons
,
K. R.
Haghi
, and
S. G.
Castleberry
,
2017
:
The structure, evolution, and dynamics of a nocturnal convective system simulated using the WRF-ARW model
.
Mon. Wea. Rev.
,
145
,
3179
3201
, https://doi.org/10.1175/MWR-D-16-0360.1.
Bleck
,
R.
, and
R.
Brummer
,
1984
:
Enhancement of remotely sensed temperature fields by wind observations from a VHF radar network
.
Mon. Wea. Rev.
,
112
,
1795
1803
, https://doi.org/10.1175/1520-0493(1984)112<1795:EORSTF>2.0.CO;2.
Bodine
,
D. J.
,
K. L.
Rasmussen
,
D. J.
Bodine
, and
K. L.
Rasmussen
,
2017
:
Evolution of mesoscale convective system organizational structure and convective line propagation
.
Mon. Wea. Rev.
,
145
,
3419
3440
, https://doi.org/10.1175/MWR-D-16-0406.1.
Bouttier
,
F.
,
2001
:
The use of profiler data at ECMWF
.
Meteor. Z.
,
10
,
497
510
, https://doi.org/10.1127/0941-2948/2001/0010-0497.
Browell
,
E. V.
,
S.
Ismail
, and
W. B.
Grant
,
1998
:
Differential Absorption Lidar (DIAL) measurements from air and space
.
Appl. Phys. B
,
67
,
399
410
, https://doi.org/10.1007/s003400050523.
Chipilski
,
H. G.
,
X.
Wang
, and
D. B.
Parsons
,
2018
:
Object-based algorithm for the identification and tracking of convective outflow boundaries in numerical models
.
Mon. Wea. Rev.
,
146
,
4179
4200
, https://doi.org/10.1175/MWR-D-18-0116.1.
Coniglio
,
M. C.
,
G. S.
Romine
,
D. D.
Turner
, and
R. D.
Torn
,
2019
:
Impacts of targeted AERI and Doppler lidar wind retrievals on short-term forecasts of the initiation and early evolution of thunderstorms
.
Mon. Wea. Rev.
,
147
,
1149
1170
, https://doi.org/10.1175/MWR-D-18-0351.1.
Cram
,
J. M.
,
M. L.
Kaplan
,
C. A.
Mattocks
, and
J. W.
Zack
,
1991
:
The use and analysis of profiler winds to derive mesoscale height and temperature fields: Simulation and real-data experiments
.
Mon. Wea. Rev.
,
119
,
1040
1056
, https://doi.org/10.1175/1520-0493(1991)119<1040:TUAAOP>2.0.CO;2.
Crook
,
N. A.
,
1986
:
The effect of ambient stratification and moisture on the motion of atmospheric undular bores
.
J. Atmos. Sci.
,
43
,
171
181
, https://doi.org/10.1175/1520-0469(1986)043<0171:TEOASA>2.0.CO;2.
Crook
,
N. A.
, and
M. W.
Moncrieff
,
1988
:
The effect of large-scale convergence on the generation and maintenance of deep moist convection
.
J. Atmos. Sci.
,
45
,
3606
3624
, https://doi.org/10.1175/1520-0469(1988)045<3606:TEOLSC>2.0.CO;2.
Crook
,
N. A.
,
R. E.
Carbone
,
M. W.
Moncrieff
, and
J. W.
Conway
,
1990
:
The generation and propagation of a nocturnal squall line. Part II: Numerical simulations
.
Mon. Wea. Rev.
,
118
,
50
66
, https://doi.org/10.1175/1520-0493(1990)118<0050:TGAPOA>2.0.CO;2.
Degelia
,
S. K.
,
X.
Wang
, and
D. J.
Stensrud
,
2019
:
An evaluation of the impact of assimilating AERI retrievals, kinematic profilers, rawinsondes, and surface observations on a forecast of a nocturnal convection initiation event during the PECAN field campaign
.
Mon. Wea. Rev.
,
147
,
2739
2764
, https://doi.org/10.1175/MWR-D-18-0423.1.
Du
,
J.
,
G.
Dimego
,
B.
Zhou
,
D.
Jovic
,
B.
Ferrier
,
B.
Yang
, and
S.
Benjamin
,
2014
:
NCEP regional ensembles: Evolving toward hourly-updated convection-allowing scale and storm-scale predictions within a unified regional modeling system. 26th Conf. on Weather Analysis and Forecasting/22nd Conf. on Numerical Weather Prediction, Atlanta, GA, Amer. Meteor. Soc., J1.4, https://ams.confex.com/ams/94Annual/webprogram/Manuscript/Paper239030/NCEP%20Regional%20Ensembles.pdf
.
Feltz
,
W. F.
,
W. L.
Smith
,
H. B.
Howell
,
R. O.
Knuteson
,
H.
Woolf
, and
H. E.
Revercomb
,
2003
:
Near-continuous profiling of temperature, moisture, and atmospheric stability using the Atmospheric Emitted Radiance Interferometer (AERI)
.
J. Appl. Meteor.
,
42
,
584
597
, https://doi.org/10.1175/1520-0450(2003)042<0584:NPOTMA>2.0.CO;2.
Ferrare
,
R.
,
A.
Nehrir
,
S.
Kooi
,
C.
Butler
, and
A.
Notari
,
2016
:
NASA DC-8 LASE data, version 1.0. UCAR/NCAR–Earth Observing Laboratory, accessed 14 April 2019, https://doi.org/10.5065/D69C6VM6
.
Fovell
,
R. G.
,
G. L.
Mullendore
, and
S.-H.
Kim
,
2006
:
Discrete propagation in numerically simulated nocturnal squall lines
.
Mon. Wea. Rev.
,
134
,
3735
3752
, https://doi.org/10.1175/MWR3268.1.
Geerts
,
B.
, and et al
,
2017
:
The 2015 Plains Elevated Convection At Night field project
.
Bull. Amer. Meteor. Soc.
,
98
,
767
786
, https://doi.org/10.1175/BAMS-D-15-00257.1.
Grasmick
,
C.
,
B.
Geerts
,
D. D.
Turner
,
Z.
Wang
, and
T. M.
Weckwerth
,
2018
:
The relation between nocturnal MCS evolution and its outflow boundaries in the stable boundary layer: An observational study of the 15 July 2015 MCS in PECAN
.
Mon. Wea. Rev.
,
146
,
3203
3226
, https://doi.org/10.1175/MWR-D-18-0169.1.
Grell
,
G. A.
,
1993
:
Prognostic evaluation of assumptions used by cumulus parameterizations
.
Mon. Wea. Rev.
,
121
,
764
787
, https://doi.org/10.1175/1520-0493(1993)121<0764:PEOAUB>2.0.CO;2.
Guo
,
Y.-R.
,
Y.-H.
Kuo
,
J.
Dudhia
,
D.
Parsons
, and
C.
Rocken
,
2000
:
Four-dimensional variational data assimilation of heterogeneous mesoscale observations for a strong convective case
.
Mon. Wea. Rev.
,
128
,
619
643
, https://doi.org/10.1175/1520-0493(2000)128<0619:FDVDAO>2.0.CO;2.
Haghi
,
K. R.
,
D. B.
Parsons
, and
B. T.
Blake
,
2015
:
Forecasting bores during PECAN 2015: A case study. 16th Conf. on Mesoscale Processes, Boston, MA, Amer. Meteor. Soc., 13.5, https://ams.confex.com/ams/16Meso/webprogram/Paper274581.html
.
Haghi
,
K. R.
,
D. B.
Parsons
,
A.
Shapiro
,
K. R.
Haghi
,
D. B.
Parsons
, and
A.
Shapiro
,
2017
:
Bores observed during IHOP_2002: The relationship of bores to the nocturnal environment
.
Mon. Wea. Rev.
,
145
,
3929
3946
, https://doi.org/10.1175/MWR-D-16-0415.1.
Haghi
,
K. R.
, and et al
,
2019
:
Bore-ing into nocturnal convection
.
Bull. Amer. Meteor. Soc.
,
100
,
1103
1121
, https://doi.org/10.1175/BAMS-D-17-0250.1.
Hartung
,
D. C.
,
J. A.
Otkin
,
R. A.
Petersen
,
D. D.
Turner
, and
W. F.
Feltz
,
2011
:
Assimilation of surface-based boundary layer profiler observations during a cool-season weather event using an observing system simulation experiment. Part II: Forecast assessment
.
Mon. Wea. Rev.
,
139
,
2327
2346
, https://doi.org/10.1175/2011MWR3623.1.
Helmus
,
J. J.
, and
S. M.
Collis
,
2016
:
The Python ARM Radar Toolkit (Py-ART), a library for working with weather radar data in the Python programming language
.
J. Open Res. Software
,
4
,
e25
, https://doi.org/10.5334/jors.119.
Hogg
,
D. C.
, and et al
,
1983a
:
An automatic profiler of the temperature, wind and humidity in the troposphere
.
J. Climate Appl. Meteor.
,
22
,
807
831
, https://doi.org/10.1175/1520-0450(1983)022<0807:AAPOTT>2.0.CO;2.
Hogg
,
D. C.
,
F. O.
Guiraud
,
J. B.
Snider
,
M. T.
Decker
, and
E. R.
Westwater
,
1983b
:
A steerable dual-channel microwave radiometer for measurement of water vapor and liquid in the troposphere
.
J. Climate Appl. Meteor.
,
22
,
789
806
, https://doi.org/10.1175/1520-0450(1983)022<0789:ASDCMR>2.0.CO;2.
Hong
,
S.-Y.
, and
J.-O. J.
Lim
,
2006
:
The WRF single-moment 6-class microphysics scheme (WSM6)
.
J. Korean Meteor. Soc.
,
42
,
129
151
.
Iacono
,
M. J.
,
J. S.
Delamere
,
E. J.
Mlawer
,
M. W.
Shephard
,
S. A.
Clough
, and
W. D.
Collins
,
2008
:
Radiative forcing by long-lived greenhouse gases: Calculations with the AER radiative transfer models
.
J. Geophys. Res.
,
113
,
D13103
, https://doi.org/10.1029/2008JD009944.
Johnson
,
A.
, and
X.
Wang
,
2017
:
Design and implementation of a GSI-based convection-allowing ensemble data assimilation and forecast system for the PECAN field experiment. Part I: Optimal configurations for nocturnal convection prediction using retrospective cases
.
Wea. Forecasting
,
32
,
289
315
, https://doi.org/10.1175/WAF-D-16-0102.1.-
Johnson
,
A.
, and
X.
Wang
,
2019
:
Multicase assessment of the impacts of horizontal and vertical grid spacing, and turbulence closure model, on subkilometer-scale simulations of atmospheric bores during PECAN
.
Mon. Wea. Rev.
,
147
,
1533
1555
, https://doi.org/10.1175/MWR-D-18-0322.1.
Johnson
,
A.
,
X.
Wang
,
J. R.
Carley
,
L. J.
Wicker
, and
C.
Karstens
,
2015
:
A comparison of multiscale GSI-based EnKF and 3DVar data assimilation using radar and conventional observations for midlatitude convective-scale precipitation forecasts
.
Mon. Wea. Rev.
,
143
,
3087
3108
, https://doi.org/10.1175/MWR-D-14-00345.1.
Johnson
,
A.
,
X.
Wang
, and
S.
Degelia
,
2017
:
Design and implementation of a GSI-based convection-allowing ensemble-based data assimilation and forecast system for the PECAN field experiment. Part II: Overview and evaluation of a real-time system
.
Wea. Forecasting
,
32
,
1227
1251
, https://doi.org/10.1175/WAF-D-16-0201.1.
Johnson
,
A.
,
X.
Wang
,
K. R.
Haghi
, and
D. B.
Parsons
,
2018
:
Evaluation of forecasts of a convectively generated bore using an intensively observed case study from PECAN
.
Mon. Wea. Rev.
,
146
,
3097
3122
, https://doi.org/10.1175/MWR-D-18-0059.1.
Jorgensen
,
D. P.
,
C. L.
Ziegler
, and
T.
Schuur
,
2016
:
NOAA P-3 1Hz flight level data, version 1.0. UCAR/NCAR–Earth Observing Laboratory, accessed 14 April 2019, https://doi.org/10.5065/D67D2SB0
.
Kain
,
J. S.
,
2004
:
The Kain–Fritsch convective parameterization: An update
.
J. Appl. Meteor.
,
43
,
170
181
, https://doi.org/10.1175/1520-0450(2004)043<0170:TKCPAU>2.0.CO;2.
Kamineni
,
R.
,
T. N.
Krishnamurti
,
R. A.
Ferrare
,
S.
Ismail
, and
E. V.
Browell
,
2003
:
Impact of high resolution water vapor cross-sectional data on hurricane forecasting
.
Geophys. Res. Lett.
,
30
,
1234
, https://doi.org/10.1029/2002GL016741.
Kamineni
,
R.
,
T. N.
Krishnamurti
,
S.
Pattnaik
,
E. V.
Browell
,
S.
Ismail
, and
R. A.
Ferrare
,
2006
:
Impact of CAMEX-4 datasets for hurricane forecasts using a global model
.
J. Atmos. Sci.
,
63
,
151
174
, https://doi.org/10.1175/JAS3588.1.
Karyampudi
,
V. M.
,
S. E.
Koch
,
C.
Chen
,
J. W.
Rottman
, and
M. L.
Kaplan
,
1995
:
The influence of the Rocky Mountains on the 13–14 April 1986 severe weather outbreak. Part II: Evolution of a prefrontal bore and its role in triggering a squall line
.
Mon. Wea. Rev.
,
123
,
1423
1446
, https://doi.org/10.1175/1520-0493(1995)123<1423:TIOTRM>2.0.CO;2.
Kawabata
,
T.
,
H.
Iwai
,
H.
Seko
,
Y.
Shoji
,
K.
Saito
,
S.
Ishii
, and
K.
Mizutani
,
2014
:
Cloud-resolving 4D-var assimilation of Doppler wind lidar data on a meso-gamma-scale convective system
.
Mon. Wea. Rev.
,
142
,
4484
4498
, https://doi.org/10.1175/MWR-D-13-00362.1.
Klein
,
P. M.
,
D. D.
Turner
,
E. N.
Smith
, and
J.
Gebauer
,
2016
:
Mobile PISA 1 OU/NSSL CLAMPS radiosonde data, version 1.0. UCAR/NCAR–Earth Observing Laboratory, accessed 14 April 2019, https://doi.org/10.5065/D6416VDH
.
Knupp
,
K.
,
2015
:
Mobile PISA 2 UAH MIPS radiosonde data, version 1.0. UCAR/NCAR–Earth Observing Laboratory, accessed 14 April 2019, https://doi.org/10.5065/D6MC8X3V
.
Knuteson
,
R. O.
, and et al
,
2004a
:
Atmospheric emitted radiance interferometer. Part I: Instrument design
.
J. Atmos. Oceanic Technol.
,
21
,
1763
1776
, https://doi.org/10.1175/JTECH-1662.1.
Knuteson
,
R. O.
, and et al
,
2004b
:
Atmospheric emitted radiance interferometer. Part II: Instrument performance
.
J. Atmos. Oceanic Technol.
,
21
,
1777
1789
, https://doi.org/10.1175/JTECH-1663.1.
Koch
,
S. E.
, and
W. L.
Clark
,
1999
:
A nonclassical cold front observed during COPS-91: Frontal structure and the process of severe storm initiation
.
J. Atmos. Sci.
,
56
,
2862
2890
, https://doi.org/10.1175/1520-0469(1999)056<2862:ANCFOD>2.0.CO;2.
Koch
,
S. E.
,
P. B.
Dorian
,
R.
Ferrare
,
S. H.
Melfi
,
W. C.
Skillman
, and
D.
Whiteman
,
1991
:
Structure of an internal bore and dissipating gravity current as revealed by Raman lidar
.
Mon. Wea. Rev.
,
119
,
857
887
, https://doi.org/10.1175/1520-0493(1991)119<0857:SOAIBA>2.0.CO;2.
Koch
,
S. E.
,
W.
Feltz
,
F.
Fabry
,
M.
Pagowski
,
B.
Geerts
,
K. M.
Bedka
,
D. O.
Miller
, and
J. W.
Wilson
,
2008a
:
Turbulent mixing processes in atmospheric bores and solitary waves deduced from profiling systems and numerical simulation
.
Mon. Wea. Rev.
,
136
,
1373
1400
, https://doi.org/10.1175/2007MWR2252.1.
Koch
,
S. E.
,
C.
Flamat
,
J. W.
Wilson
,
B. M.
Gentry
, and
B. D.
Jamison
,
2008b
:
An atmospheric soliton observed with Doppler radar, differential absorption lidar, and a molecular Doppler lidar
.
J. Atmos. Oceanic Technol.
,
25
,
1267
1287
, https://doi.org/10.1175/2007JTECHA951.1.
Kuo
,
Y.-H.
,
E. G.
Donall
, and
M. A.
Shapiro
,
1987
:
Feasibility of short-range numerical weather prediction using observations from a network of profilers
.
Mon. Wea. Rev.
,
115
,
2402
2427
, https://doi.org/10.1175/1520-0493(1987)115<2402:FOSRNW>2.0.CO;2.
Liu
,
C.
, and
M. W.
Moncrieff
,
1996
:
A numerical study of the effects of ambient flow and shear on density currents
.
Mon. Wea. Rev.
,
124
,
2282
2303
, https://doi.org/10.1175/1520-0493(1996)124<2282:ANSOTE>2.0.CO;2.
Loehrer
,
S. M.
,
T. A.
Edmands
, and
J. A.
Moore
,
1996
:
TOGA COARE upper-air sounding data archive development and quality control procedures
.
Bull. Amer. Meteor. Soc.
,
77
,
2651
2672
, https://doi.org/10.1175/1520-0477(1996)077<2651:TCUASD>2.0.CO;2.
Melfi
,
S. H.
, and
D.
Whiteman
,
1985
:
Observation of lower-atmospheric moisture structure and its evolution using a Raman lidar
.
Bull. Amer. Meteor. Soc.
,
66
,
1288
1292
, https://doi.org/10.1175/1520-0477(1985)066<1288:OOLAMS>2.0.CO;2.
Melfi
,
S. H.
,
D.
Whiteman
, and
R.
Ferrare
,
1989
:
Observation of atmospheric fronts using Raman lidar moisture measurements
.
J. Appl. Meteor.
,
28
,
789
806
, https://doi.org/10.1175/1520-0450(1989)028<0789:OOAFUR>2.0.CO;2.
Mlawer
,
E. J.
,
S. J.
Taubman
,
P. D.
Brown
,
M. J.
Iacono
, and
S. A.
Clough
,
1997
:
Radiative transfer for inhomogeneous atmospheres: RRTM, a validated correlated-k model for the longwave
.
J. Geophys. Res.
,
102
,
16 663
16 682
, https://doi.org/10.1029/97JD00237.
Moore
,
A. S.
, and et al
,
1997
:
Development of the Lidar Atmospheric Sensing Experiment (LASE)—An advanced airborne dial instrument. Advances in Atmospheric Remote Sensing with Lidar, A. Ansmann et al., Eds, Springer, 281–288
.
Nakanishi
,
M.
, and
H.
Niino
,
2006
:
An improved Mellor-Yamada level-3 model: Its numerical stability and application to a regional prediction of advection fog
.
Bound.-Layer Meteor.
,
119
,
397
407
, https://doi.org/10.1007/s10546-005-9030-8.
NRC
,
2009
:
Observing Weather and Climate from the Ground Up. National Academies Press, 250 pp
.
NRC
,
2010
:
When Weather Matters: Science and Service to Meet Critical Societal Needs. National Academies Press, 198 pp
.
NRC
,
2012
:
Weather Services for the Nation: Becoming Second to None. National Academies Press, 86 pp
.
Otkin
,
J. A.
,
D. C.
Hartung
,
D. D.
Turner
,
R. A.
Petersen
,
W. F.
Feltz
, and
E.
Janzon
,
2011
:
Assimilation of surface-based boundary layer profiler observations during a cool-season weather event using an observing system simulation experiment. Part I: Analysis impact
.
Mon. Wea. Rev.
,
139
,
2309
2326
, https://doi.org/10.1175/2011MWR3622.1.
Parker
,
M. D.
,
2008
:
Response of simulated squall lines to low-level cooling
.
J. Atmos. Sci.
,
65
,
1323
1341
, https://doi.org/10.1175/2007JAS2507.1.
Parsons
,
D. B.
,
K. R.
Haghi
,
K. T.
Halbert
,
B.
Elmer
, and
J.
Wang
,
2019
:
The potential role of atmospheric bores and gravity waves in the initiation and maintenance of nocturnal convection over the Southern Great Plains
.
J. Atmos. Sci.
,
76
,
43
68
, https://doi.org/10.1175/JAS-D-17-0172.1.
Revercomb
,
H. E.
,
H.
Buijs
,
H. B.
Howell
,
D. D.
LaPorte
,
W. L.
Smith
, and
L. A.
Sromovsky
,
1988
:
Radiometric calibration of IR Fourier transform spectrometers: Solution to a problem with the high-resolution interferometer sounder
.
Appl. Opt.
,
27
,
3210
3218
, https://doi.org/10.1364/AO.27.003210.
Rottman
,
J.
, and
J.
Simpson
,
1989
:
The formation of internal bores in the atmosphere: A laboratory model
.
Quart. J. Roy. Meteor. Soc.
,
115
,
941
963
, https://doi.org/10.1002/qj.49711548809.
Schwartz
,
C. S.
, and et al
,
2010
:
Toward improved convection-allowing ensembles: Model physics sensitivities and optimizing probabilistic guidance with small ensemble membership
.
Wea. Forecasting
,
25
,
263
280
, https://doi.org/10.1175/2009WAF2222267.1.
Skamarock
,
W. C.
, and et al
,
2008
:
A description of the Advanced Research WRF version 3. NCAR Tech. Note NCAR/TN-475+STR, 113 pp., https://doi.org/10.5065/D68S4MVH
.
Smith
,
T. L.
, and
S. G.
Benjamin
,
1993
:
Impact of network wind profiler data on a 3-h data assimilation system
.
Bull. Amer. Meteor. Soc.
,
74
,
801
807
, https://doi.org/10.1175/1520-0477(1993)074<0801:IONWPD>2.0.CO;2.
Tewari
,
M.
, and et al
,
2004
:
Implementation and verification of the unified Noah land surface model in the WRF Model. 20th Conf. on Weather Analysis and Forecasting/16th Conf. on Numerical Weather Prediction, Seattle, WA, Amer. Meteor. Soc., 14.2A, https://ams.confex.com/ams/84Annual/techprogram/paper_69061.htm
.
Thompson
,
G.
,
P. R.
Field
,
R. M.
Rasmussen
, and
W. D.
Hall
,
2008
:
Explicit forecasts of winter precipitation using an improved bulk microphysics scheme. Part II: Implementation of a new snow parameterization
.
Mon. Wea. Rev.
,
136
,
5095
5115
, https://doi.org/10.1175/2008MWR2387.1.
Turner
,
D. D.
,
2016a
:
FP4 AERIoe thermodynamic profile retrieval data, version 2.0. UCAR/NCAR–Earth Observing Laboratory, accessed 14 April 2019, https://doi.org/10.5065/D6PN93T3
.
Turner
,
D. D.
,
2016b
:
MP1 OU/NSSL CLAMPS Doppler lidar VAD wind data, version 1.0. UCAR/NCAR–Earth Observing Laboratory, accessed 14 April 2019,
https://doi.org/10.5065/D6BR8QJH.
Turner
,
D. D.
,
2018
:
MP1 OU/NSSL CLAMPS AERIoe thermodynamic profile retrieval data, version 1.2. UCAR/NCAR–Earth Observing Laboratory, accessed 14 April 2019, https://doi.org/10.5065/D6VQ312C
.
Turner
,
D. D.
, and
U.
Löhnert
,
2014
:
Information content and uncertainties in thermodynamic profiles and liquid cloud properties retrieved from the ground-based atmospheric emitted radiance interferometer (AERI)
.
J. Appl. Meteor. Climatol.
,
53
,
752
771
, https://doi.org/10.1175/JAMC-D-13-0126.1.
UCAR/NCAR
,
2015
:
FP4 NCAR/EOL 915 MHz Profiler NIMA consensus winds and moments, version 1.0. UCAR/NCAR–Earth Observing Laboratory, accessed 14 April 2019, https://doi.org/10.5065/D6RV0KXH
.
UCAR/NCAR
,
2016a
:
FP4 NCAR/EOL QC soundings, version 2.0. UCAR/NCAR–Earth Observing Laboratory, accessed 14 April 2019, https://doi.org/10.5065/D63776XH
.
UCAR/NCAR
,
2016b
:
MP4 NCAR/EOL QC soundings, version 2.0. UCAR/NCAR–Earth Observing Laboratory, accessed 14 April 2019, https://doi.org/10.5065/D6707ZNV
.
UCAR/NCAR
,
2016c
:
FP3 FP4 FP5 QC 5 min surface data, tilt corrected, version 1.0. UCAR/NCAR–Earth Observing Laboratory, accessed 14 April 2019, https://doi.org/10.5065/D6BZ645V
.
UCAR/NCAR
,
2016d
:
MP4 NCAR/EOL MISS 915 MHz Profiler 30 minute consensus winds and moments and surface meteorology data, version 1.0. UCAR/NCAR–Earth Observing Laboratory, accessed 14 April 2019, https://doi.org/10.5065/D6RJ4GPJ
.
Wagner
,
T. J.
,
E.
Olson
,
N.
Smith
, and
W.
Feltz
,
2016a
:
MP3 University of Wisconsin SPARC AERIoe thermodynamic profile data, version 1.0. UCAR/NCAR–Earth Observing Laboratory, accessed 14 April 2019, https://doi.org/10.5065/D60Z71HC
.
Wagner
,
T. J.
,
E.
Olson
,
N.
Smith
, and
W.
Feltz
,
2016b
:
MP3 University of Wisconsin SPARC surface meteorological data, version 1.0. UCAR/NCAR–Earth Observing Laboratory, accessed 14 April 2019, https://doi.org/10.5065/D6N014XZ
.
Wagner
,
T. J.
,
D. D.
Turner
, and
R.
Newsom
,
2016c
:
MP3 University of Wisconsin SPARC Doppler lidar VAD wind data, version 2.0. UCAR/NCAR–Earth Observing Laboratory, accessed 14 April 2019, https://doi.org/10.5065/D6V9869B
.
Wang
,
Y.
, and
X.
Wang
,
2017
:
Direct assimilation of radar reflectivity without tangent linear and adjoint of the nonlinear observation operator in the GSI-based EnVar system: Methodology and experiment with the 8 May 2003 Oklahoma City tornadic supercell
.
Mon. Wea. Rev.
,
145
,
1447
1471
, https://doi.org/10.1175/MWR-D-16-0231.1.
Wang
,
Z.
, and et al
,
2016
:
University of Wyoming king air compact Raman lidar data, version 1.0. UCAR/NCAR–Earth Observing Laboratory, accessed 14 April 2019, https://doi.org/10.5065/D6MS3R0P
.
Waugh
,
S.
, and
C.
Ziegler
,
2017
:
NSSL mobile mesonet data, version 1.1. UCAR/NCAR–Earth Observing Laboratory, accessed 23 December 2019, https://doi.org/10.5065/D64M92RG
.
Weber
,
B. L.
, and et al
,
1990
:
Preliminary evaluation of the first NOAA demonstration network wind profiler
.
J. Atmos. Oceanic Technol.
,
7
,
909
918
, https://doi.org/10.1175/1520-0426(1990)007<0909:PEOTFN>2.0.CO;2.
Weckwerth
,
T. M.
,
K. J.
Weber
,
D. D.
Turner
, and
S. M.
Spuler
,
2016
:
Validation of a water vapor micropulse Differential Absorption Lidar (DIAL)
.
J. Atmos. Oceanic Technol.
,
33
,
2353
2372
, https://doi.org/10.1175/JTECH-D-16-0119.1.
Wei
,
M.
,
Z.
Toth
,
R.
Wobus
, and
Y.
Zhu
,
2008
:
Initial perturbations based on the Ensemble Transform (ET) technique in the NCEP global operational forecast system
.
Tellus
,
60A
,
62
79
, https://doi.org/10.1111/j.1600-0870.2007.00273.x.
Weisman
,
M. L.
,
C.
Davis
,
W.
Wang
,
K. W.
Manning
, and
J. B.
Klemp
,
2008
:
Experiences with 0–36-h explicit convective forecasts with the WRF-ARW model
.
Wea. Forecasting
,
23
,
407
437
, https://doi.org/10.1175/2007WAF2007005.1.
Westwater
,
E. R.
,
W. B.
Sweezy
,
L. M.
McMillin
, and
C.
Dean
,
1984
:
Determination of atmospheric temperature profiles from a statistical combination of ground-based profiler and operational NOAA 6/7 satellite retrievals
.
J. Climate Appl. Meteor.
,
23
,
689
703
, https://doi.org/10.1175/1520-0450(1984)023<0689:DOATPF>2.0.CO;2.
White
,
B. L.
, and
K. R.
Helfrich
,
2012
:
A general description of a gravity current front propagating in a two-layer stratified fluid
.
J. Fluid Mech.
,
711
,
545
575
, https://doi.org/10.1017/jfm.2012.409.
Whiteman
,
D. N.
, and et al
,
2006
:
Raman lidar measurements during the International H2O Project. Part II: Case studies
.
J. Atmos. Oceanic Technol.
,
23
,
170
183
, https://doi.org/10.1175/JTECH1839.1.
Zhang
,
J.
, and et al
,
2016
:
Multi-Radar Multi-Sensor (MRMS) quantitative precipitation estimation: Initial operating capabilities
.
Bull. Amer. Meteor. Soc.
,
97
,
621
638
, https://doi.org/10.1175/BAMS-D-14-00174.1.
Zhang
,
L.
, and
Z.
Pu
,
2011
:
Four-dimensional assimilation of multitime wind profiles over a single station and numerical simulation of a mesoscale convective system observed during IHOP_2002
.
Mon. Wea. Rev.
,
139
,
3369
3388
, https://doi.org/10.1175/2011MWR3569.1.
Ziegler
,
C. L.
,
M.
Coniglio
,
M.
Parker
,
R. S.
Schumacher
, and
K.
Young
,
2016
:
CSU/NCSU/NSSL MGAUS radiosonde data, version 3.0. UCAR/NCAR–Earth Observing Laboratory, accessed 14 April 2019, https://doi.org/10.5065/D6W66HXN
.
For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Footnotes

1

Throughout this paper the term radiosonde would be used in lieu of rawinsonde to denote balloon launches with GPS readings.

2

Localization (and observation error) values were tuned to maximize the skill of the convective forecasts.

3

Examination of the raw wind profiler data revealed large deviations in wind direction when compared to the nearby mobile radiosonde units.

4

Mean inversion wind is defined as the wind averaged over the depth of the SBL.

5

MG stands for a Mobile GPS Advanced Upper-Air Sounding System.

6

The estimation of the mean average errors (MAEs) is a two-step process. First, the ensemble mean error is computed at a particular reference point based on all available ensemble members. Then, the ensemble mean errors are averaged over all reference points falling within a particular neighborhood radius.