1. Introduction
Past studies have shown that changes occur thermodynamically and kinematically in those convective boundary layers that make a transition from turbulent daytime patterns to stable nocturnal patterns over land surfaces. During this time period, surface temperature and wind speed decrease (Mahrt 1981; Acevedo and Fitzjarrald 2001), water vapor increases, and surface heat flux values make a transition from positive to negative. This study aims to examine these changes using data from ground-based remote sensing instruments (Doppler sodar in particular) and surface instruments to develop a more complete understanding of the afternoon-to-evening transition (AET).
Past definitions of the AET have been based on a variety of measurements and atmospheric conditions. The most-popular definition is the onset of negative surface heat flux (Caughey et al. 1979; Grant 1997; Beare et al. 2006). According to this definition, the AET is an almost-instantaneous local event that occurs when the daytime positive surface heat flux switches to the nighttime negative heat flux. This is a convenient definition because it is easy to determine with surface flux measurements. Edwards et al. (2006) expand on this definition by adding that shortwave heating decreases during the AET period even in the presence of a positive surface heat flux. This modification extends the AET to earlier in the day but still does not account for any changes that occur after the heat flux changes sign. Nieuwstadt and Brost (1986) assumed that the AET occurs when upward surface sensible heat flux ceases. We find that this definition does not encompass all physical processes of the transition that occur over a deeper layer and extended time period.
Other studies have proposed different definitions of the AET. Mahrt (1981) stated that the beginning occurred when the low-level winds (below 50–100 m) began to decelerate about 2 h before sunset and that the ending occurred when flow at all levels in the nocturnal boundary layer rotated toward high pressure. Following this kinematic definition, the AET takes 4 or 5 h, which is very different from the definitions used in other studies. Although this definition takes into account the conditions above the surface, it does not account for any changes occurring near the surface. The stable boundary layer forms upward from the surface, and therefore to discount any changes at the surface could potentially miss part of the transition. Another problem is that the winds in all geographic regions have not been found to rotate in the same direction. Caughey et al. (1979) and Acevedo and Fitzjarrald (2001) found no preferential direction in the rotation of the winds. Stull (1988), however, states that winds often veer with height after formation of the stable boundary layer because of the inertial oscillation.
Acevedo and Fitzjarrald (2001) defined the beginning of the AET period as when a sharp decrease in the spatial temperature difference occurred and defined the ending as when the spatial standard deviation of temperature reached a maximum. This definition is convenient because it only depends on temperature measurements. It was developed in a region with variable terrain height (Albany, New York), however.
Another definition was based on measurements from a wind profiler (Grimsdell and Angevine 2002). Their definition stated that the AET occurred when the convective boundary layer height, as defined by 915-MHz profiler measurements of return power, began to decrease and there was a sharp decrease in spectral width in a given layer. An updated definition is provided in Angevine (2008), who defined the boundary layer transition as the decrease of turbulence near the top of the daytime convective boundary layer. This definition is similar to the spectral-width definition from the previous paper. No attention is paid to changes occurring near the surface in either paper. Also, if turbulence is weak throughout the day, then using only turbulence to determine the onset of the boundary layer transition will not work.
Mahrt (1981) found that winds typically decelerate in the lowest 50–100 m about 2 h before sunset. Our findings reveal that inconsistent information about the timing of wind speed reductions during the early transition period prevents development of a generalized pattern for the behavior of wind speeds during the transition period, but a deceleration of the low-level (≤100 m AGL) winds is generally expected because of a reduction in downward turbulent transport of momentum associated with increased stability. Winds at higher levels begin to accelerate during this time period, consistent with the formation of a nocturnal jet.
This study examines the changes that take place during the AET by using sodar, radar wind profiler, and surface data to assess the general behavior of the atmospheric characteristics during the transition period. Using these data, we examine the evolution of the AET for 30 cases, with a focus on 5 August 2007, to determine whether observations fit with expectations from previous definitions. This information will then be used in a future study to determine the processes associated with convective initiation and intensification during the transition time period. Section 2 discusses the data and methods used in this study. The results of the study are given in section 3, and discussion is given in section 4. The conclusions are in section 5.
2. Data and method
Mobile Integrated Profiling System sensors and their measurement characteristics (Karan 2007).
Inhomogeneities in temperature produce acoustic backscattering. This fact means that as the heat flux decreases during the transition period the temperature within the lower boundary layer becomes more homogeneous and the acoustic backscatter decreases. We have found that the relative minimum in sodar backscatter intensity and height coverage during this period provides an important measure of the transition in the current study. Therefore, the sodar backscatter [signal-to-noise ratio (SNR)] can be used to visualize the transition from convective structures to stable structures within the boundary layer.
The sodar vertical velocity w profiles are also useful in determining the structure of the boundary layer. The convective afternoon boundary layer is associated with frequent plumes, which are marked by pockets of upward vertical motion. They are also present as inhomogeneities in the SNR profiles. As the boundary layer stabilizes, both w and SNR become more homogeneous.
The sodar wind profiles are examined to determine the behavior of the winds in the lowest part of the atmosphere during the transition. Wind speed and direction are calculated every 15 min at 25-m vertical resolution, starting at 50 m AGL. An average of the wind speeds at 100, 125, and 150 m was examined to determine temporal patterns above the surface during the transition. These three levels were chosen (and averaged) to increase the sample size and thereby to reduce the noise in the temporal trend of winds above the surface layer.
The 915-MHz profiler is used to examine the winds and SNR throughout the boundary layer. The sodar is able to capture the lowest layers of the atmosphere, whereas the profiler is able to gather data up to 3–5 km AGL. A 915-MHz profiler was also previously used by Grimsdell and Angevine (2002) to examine the boundary layer transition. This study will further examine the boundary layer by combining results from the 915-MHz profiler with the observations from the other instrumentation.
The surface data used in this study include 2-m temperature (accuracy of 0.2 K) and relative humidity (accuracy of 2%) derived from a Vaisala, Inc., HMP-45C sensor, 10-m temperature (Campbell Scientific, Inc., 107-L Fast Response Sensor, with an accuracy of 0.2 K), and 10-m wind speed (from an R.M. Young Co. 051303-L wind monitor). Time series of these data are examined to find characteristics that can be linked to the transition. These results are then used in conjunction with the sodar data to find a general pattern for the clear-air boundary layer.
The variation in V10m is also examined to identify the kinematic changes during the AET period. After all the wind speed time series have been analyzed, an average wind speed pattern for all cases is identified to help to refine the characteristics of the transition time period.
The data from 21 mostly clear days (cumulus cloud cover of <60% during midafternoon) from June through August 2007 and 9 clear days from October and November 2008 in Huntsville were analyzed to examine the characteristics of the AET. For the sake of this study, clear days are defined as any day without extensive cloud cover or precipitation, meaning that days with scattered clouds were also included. The percentage of spatial cloud cover is assumed to be equal to the percentage of temporal cloud coverage as determined using measurements from a ceilometer (Vaisala CT25K) and an Eppley Laboratory, Inc., pyranometer, both collocated with the sodar and surface instrumentation.
After clear days or days with only scattered clouds were identified, surface data from those days were analyzed during the period from 1400 to 2100 LST (2000–0300 UTC). The sodar began operating at 1600 LST for most days, and therefore sodar data were analyzed from 1600 to 2100 LST. Analysis of temperature, mixing ratio, wind speed, and sodar data was conducted to determine the general characteristics of the clear-air boundary layer transition as well as to find a definition of the boundary layer transition using the sodar and surface instruments.
3. Results
The data from the 30 cases were analyzed and were divided into two categories on the basis of season. The 21 summer cases were found to have different characteristics from the 9 autumn cases. A case study from 5 August 2007 is discussed in detail to illustrate the specific characteristics of the atmospheric boundary layer (ABL) transition for a single day. Then, comprehensive results from the summer cases as well as nine autumn cases are discussed.
a. 5 August 2007
Although the 5 August 2007 case involved the presence of clouds, as indicated by both the ceilometer and the pyranometer, the AET is still effectively illustrated in the data from the surface instruments, sodar, and 915-MHz wind profiler. According to the data from these instruments, the cumulus cloud cover decreased during the transition time period. Around 1800 LST, only widely scattered altocumulus clouds were present near 4.8 km AGL. Most of the days included in this study experienced decreasing cumulus cloud cover as the AET progressed.
On 5 August, a low pressure center was located over eastern Nebraska and a high pressure center was over western Virginia, leading to dominant high pressure over Huntsville. The maximum afternoon temperature and dewpoint were 36° and 20°C, respectively. The geostrophic wind was westerly at approximately 3 m s−1. During the 1400–1500 LST period, the ceilometer-derived cumulus cloud cover was approximately 40%, and cloud-base height was approximately 1800 m AGL (Fig. 1), and the CBL depth was 2 km (from the 915-MHz profiler vertical velocity and SNR data; Fig. 2). Sunset on 5 August occurred at 1845 LST.
Cloud-base height as determined from ceilometer measurements from 5 Aug 2007 (vertical black line indicates time of sunset).
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
Cloud-base height as determined from ceilometer measurements from 5 Aug 2007 (vertical black line indicates time of sunset).
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
Cloud-base height as determined from ceilometer measurements from 5 Aug 2007 (vertical black line indicates time of sunset).
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
915-MHz profiler (top) SNR and (bottom) vertical velocity from 5 Aug 2007 (vertical black line indicates time of sunset).
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
915-MHz profiler (top) SNR and (bottom) vertical velocity from 5 Aug 2007 (vertical black line indicates time of sunset).
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
915-MHz profiler (top) SNR and (bottom) vertical velocity from 5 Aug 2007 (vertical black line indicates time of sunset).
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
The sodar data can be used to illustrate further the AET. While the afternoon convective boundary layer is present, surface-layer plumes, as represented by columns of enhanced backscatter, are indicated in the sodar SNR time–height section in Fig. 3. As the transition occurs, the fluctuations in SNR decrease and a more continuous, horizontally stratified layer becomes visible. On 5 August, the higher-frequency fluctuations in sodar SNR decreased by 1800 LST at the time of minimum sodar backscatter and the associated transition 45 min before sunset. The stratified layer associated with the incipient nocturnal boundary layer (NBL) can be seen by 1840 LST (5 min before sunset). Following this change, the stratified layer deepened and strengthened as the NBL developed.
Sodar (top) backscatter and (bottom) vertical velocity from 5 Aug 2007 (vertical black line indicates time of sunset).
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
Sodar (top) backscatter and (bottom) vertical velocity from 5 Aug 2007 (vertical black line indicates time of sunset).
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
Sodar (top) backscatter and (bottom) vertical velocity from 5 Aug 2007 (vertical black line indicates time of sunset).
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
The sodar vertical velocity profiles reveal updrafts and downdrafts associated with the fluctuations in sodar return power during CBL conditions (Fig. 3). The ABL transition is closely associated with the dissipation of these significant vertical motion perturbations. This change occurred 45 min after the transition at approximately 1845 LST (at sunset) on 5 August and corresponds to the formation of the NBL.
The sodar SNR and vertical velocity structures are similar to the 915-MHz profiler SNR and vertical velocity (Fig. 2) but on a much shallower scale. The 915-MHz profiler reveals significant thermals during the convective time period. Large fluctuations near the top of the convective boundary layer represent cumulus clouds forming above the mixed layer. These thermals are also visible on the vertical velocity time–height series with areas of upward and downward motion present prior to the AET. The last significant thermal was detected just before 1700 LST (105 min before sunset). After the AET, the thermals are no longer visible on the SNR or vertical velocity time–height series. The 915-MHz SNR shows less turbulence near the top of the boundary layer, and it decays to small values.
Figures 4a and 4c show the 915-MHz profiler average SNR from the lowest three range gates (106, 163, and 220 m) at 55-s temporal resolution and the 150-m sodar SNR at 25-s temporal resolution, respectively. Because of a software error, 915-MHz data were not archived between 1730 and 1800 LST. However, the 915-MHz SNR data show the difference between the more-convective signature and the more-stable signature after the transition. Before the data gap, the SNR values are highly variable. After the gap, the SNR data become more consistent, with smaller fluctuations. The sodar SNR shows a lower relative variation in the signal, but it does show the relative minimum at the transition time (~1750 LST; 55 min before sunset). The vertical velocities w from both instruments show the transition from highly variable to much lower fluctuations after 1800 LST (Figs. 4b,d).
Time series of data from various sensors during the boundary layer transition on 5 Aug 2007 (vertical black line indicates time of sunset).
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
Time series of data from various sensors during the boundary layer transition on 5 Aug 2007 (vertical black line indicates time of sunset).
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
Time series of data from various sensors during the boundary layer transition on 5 Aug 2007 (vertical black line indicates time of sunset).
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
The 2-m temperature (T2m; Fig. 4e) and 10-m temperature (T10m; not shown) time series from 5 August show a nearly steady value of 36° and 35°C, respectively, until an initial decrease near 1700 LST (105 min before sunset). The T2m − T10m difference decreased to a near-zero value at approximately 2020 LST (95 min after sunset; Fig. 4f). The formation of an inversion is often associated with the ABL transition but is not always present for the summer cases, as indicated here and discussed further in the following section. The use of T2m − T10m as a proxy for surface heat flux therefore becomes tenuous in this case because the temperature difference did not change sign, although it was near zero. A noticeable feature of the temperatures at both levels and the temperature-difference values is that the variability of T2m and T10m data generally decreased with time after 1525 LST (200 min before sunset).
Another surface characteristic showing a significant change during the AET is the mixing ratio (Fig. 4g). As turbulent mixing (a sink of low-level water vapor) decreases, the low-level moisture typically increases. The mixing ratio increased monotonically after 1720 LST (85 min before sunset), about 20 min after the initial temperature decrease, placing the initiation of this increase during the transition. The mixing ratio variances decrease in much the same way as the temperature variances decrease after 1500 LST (225 min before sunset).
The 10-m wind speed V10m and variance 
Winds aloft derived from the sodar and 915-MHz profiler show a veering of the winds from southwesterly at 1600 LST (165 min before sunset) to a more westerly direction by 1730 LST (75 min before sunset). The 915-MHz profiler shows a similar veering pattern from southwesterly to westerly winds below 500 m. Above 500 m, the winds remained westerly throughout the time period. The 915-MHz profiler wind speed at 1 km AGL decreased during the formation of the stable boundary layer at 1800 LST (45 min before sunset) and then increased again by 2100 LST (135 min after sunset; Fig. 4j). Sodar average wind speeds between 100 and 150 m AGL increased from 4 m s−1 at 1600 LST (165 min before sunset) to 7 m s−1 at 2000 LST (75 min after sunset; Fig. 4k). The increase within this height interval is not typical for all cases, as will be discussed in the next section.
The preceding combination of the datasets illustrates that the transition appears to have occurred over a broad time period (80 min) between 1700 and 1820 LST (105–25 min before sunset). The transition started with the decrease in temperature and ended with the decrease in 10-m wind speed. During the AET period, there was also an increase in mixing ratio at 2 m and in sodar-derived winds aloft and a decrease in 10-m wind speed, wind speed variance, and temperature variance, as well as decreases in sodar return power and vertical velocity magnitudes. The sodar SNR shows a progression from a convective structure with active thermals to a minimum in backscatter as the surface heat flux reaches zero, followed by the appearance of a more-stratified structure associated with the formation of the NBL.
b. June–August cases
Additional cases from the summer of 2007 show changes during the transition period that are similar to those shown in the 5 August case above. Those changes do not necessarily occur in the same order, however. Changes at the surface tend to occur before changes in sodar backscatter, which are related to temperature fluctuations above the surface. The same general conditions are present for most of the summer cases. The average 10-m wind speed for the 21 summer cases was approximately 3 m s−1 prior to the transition (Fig. 5). There were generally no fronts in the area. The cloud conditions for these days ranged from clear to approximately 50% cumulus cloud cover. These common characteristics contributed to the relatively consistent pattern for the ABL transition.
Distribution of average afternoon 10-m wind speeds for the 21 summer cases (black) and the 9 autumn cases (gray).
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
Distribution of average afternoon 10-m wind speeds for the 21 summer cases (black) and the 9 autumn cases (gray).
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
Distribution of average afternoon 10-m wind speeds for the 21 summer cases (black) and the 9 autumn cases (gray).
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
The first characteristic of the ABL transition is generally the decrease of wind speed variance, as seen on 5 August. This was the first variable to change for 14 of the 21 cases (Fig. 6). The earliest this change occurred relative to sunset was on 21 July at 1515 LST (222 min before sunset), and the latest was on 11 August at 1900 LST (21 min after sunset). The decrease in wind speed occurs from 55 min before the decrease in wind speed variance to 165 min after. In cases in which the wind speed decreased much later than the wind speed variance, the decrease in wind speed was the last transition characteristic to appear and the wind speed variance decrease was the first.
Box plot illustrating the distribution of times relative to sunset for the transition during June–August. The box represents the middle 50% of values, with each whisker representing 25% of the values.
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
Box plot illustrating the distribution of times relative to sunset for the transition during June–August. The box represents the middle 50% of values, with each whisker representing 25% of the values.
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
Box plot illustrating the distribution of times relative to sunset for the transition during June–August. The box represents the middle 50% of values, with each whisker representing 25% of the values.
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
The temperature decrease at 2 and 10 m typically occurs shortly after the decrease in variance of wind speed. In four cases, the temperature decrease occurred before the decrease in wind speed variance. The temperature is relatively steady during the hours leading up to the decrease. A temperature inversion, which is strong evidence that the surface heat flux has changed signs, formed in 15 of the 21 cases. Because the inversion is not always present, however, the presence of an inversion cannot be used alone to define the transition. If an inversion forms, it is typically one of the last elements of the transition to fall into place. The formation of an inversion appears to be related to wind speed and mixing ratio. Cases without an inversion tend to have higher wind speeds and higher mixing ratios; higher wind speeds and mixing ratios do not mean that an inversion will not form, however.
Even if no inversion forms, a significant decrease in the temperature difference is observed. This decrease can be considered to be a part of the transition even in the absence of a temperature inversion. The temperature difference variance decreases for all cases regardless of the presence of a temperature inversion. The beginning of a relative minimum in temperature difference variance can also be considered to be part of the ABL transition.
The mixing ratio pattern seen on 5 August is a typical pattern for the ABL transition period. During the transition a steady increase begins and the variability decreases. In most cases, the mixing ratio increase occurs just after the temperature decrease but before the temperature inversion forms, as on 5 August, implying a reduction in mixing and TKE. This again indicates that the increase in mixing ratio can be considered to be part of the ABL transition.
The sodar signature for the ABL transition is relatively consistent for all of the clear summer cases. The sodar backscatter experiences fluctuations corresponding to thermals during the convective boundary layer period. Sometime after 1800 LST, the fluctuations disappear and a low-level stable layer forms. The relative minimum in sodar return power just prior to the development of the stable layer is the first sign of the transition in the sodar data. This change also shows a good correlation with the inversion formation, consistent with the physics of sodar backscatter. As time progresses, the stable layer and associated sodar backscatter deepen and the SNR magnitude in that layer increases.
The sodar vertical velocity profiles also show the transition to the more-stable boundary layer. The areas of relatively strong upward or downward motion present during the afternoon dissipate during the transition and are no longer present after the NBL forms. The vertical velocity dissipation often corresponds to the disappearance of the thermal plumes in the backscattered power time series.
An examination of the sodar wind profiles for the 21 summer cases reveals that there is no consistent pattern during the transition. The winds do not show a tendency to rotate in any particular direction, and there does not appear to be a consistent pattern for change in the wind speeds. An examination of the 100–150-m wind speeds reveals that although the winds decrease near the surface the winds aloft do not follow the same pattern. No consistent overall pattern can be seen, even though almost 50% of the cases (e.g., 5 August) do have wind speed increases within the 100–150-m layer.
The 915-MHz profiler data for each summer case are similar to the 5 August case in that the profiler shows a prominent reduction in deep thermals leading up to the AET. The depth of the convective boundary layer and the overall pattern of the SNR and vertical velocity profiles differ for each case, however. The most common characteristic is a reduction in turbulence associated with the thermals. Using the 915-MHz profiler in this way allows for examination of the transition over the entire CBL depth, rather than being limited to the depth of the sodar return. These results can also be compared with those of Angevine (2008).
The basic pattern for the summer afternoon–evening AET is for the 10-m wind speed variance to decrease, indicating a decrease in TKE, followed by a decrease in the low-level temperatures (Table 2; Fig. 6). Toward the end of the transition, the sodar return power shows a minimum in both height coverage and return power, followed by a gradual increase that marks the development of the nocturnal stable layer. An increase in mixing ratio and decreases in 10-m wind speed and temperature difference also occur during the transition. These general characteristics can be found during the transition for almost all of the summer cases.
Transition times (in minutes) relative to sunset for summer cases (negative values are before sunset; positive values are after sunset).
c. October and November cases
Nine consecutive days from 29 October to 6 November 2008 were included in this study to examine characteristics of the atmospheric ABL transition during the cooler and drier autumn season. The analysis reveals that some of the same basic characteristics are present during the colder months but that the order in which these changes occur and the intensity of the changes are different.
The beginning of the transition is no longer indicated by the decrease in the wind speed variance. While this change still occurs, it is not the dominant first feature (Fig. 7). The beginning of the transition can be indicated by a variety of factors. The temperature time series show that the beginning temperatures are much lower than for the previous cases, as would be expected (e.g., 2 November; Fig. 8a). The temperature decreases also start closer to sunset (1651 LST, or 2251 UTC; 3 min before sunset) and are much more dramatic. The temperature inversions are also much stronger. A temperature inversion formed for each of the nine cases included in the study. After the inversion, the magnitudes of the temperature differences are often larger than the temperature differences before the inversion. Although these changes are more intense than the changes during the summer months, the fundamental pattern is the same.
As in Fig. 6, but during October and November.
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
As in Fig. 6, but during October and November.
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
As in Fig. 6, but during October and November.
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
Time series of data during the boundary layer transition on 2 Nov 2008 (vertical black line indicates time of sunset).
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
Time series of data during the boundary layer transition on 2 Nov 2008 (vertical black line indicates time of sunset).
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
Time series of data during the boundary layer transition on 2 Nov 2008 (vertical black line indicates time of sunset).
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
The mixing ratios for the autumn cases increase during the transition (Fig. 8b), as they did during the summer months. The mixing ratio values typically start with lower values than the starting summer values and then increase at a greater rate than the summer mixing ratios. The mixing ratio variances are similar to the variances in temperature. Large increases occur toward the end of the transition at which point the increases in mixing ratio tend to be the greatest.
The 10-m wind speeds are generally lower for these autumn cases than for the summer cases (Fig. 8c). After about 1700 LST (sunset), the wind speeds often decrease to near-calm conditions, with some small fluctuations. The wind speed variances usually do not exhibit the same increases as do the variances in moisture and temperature. The typical decrease in wind speed variance occurs, with values reaching zero as the wind speeds reach zero.
The sodar runs continuously during weekends, which allows for better examination of the ABL transition (Fig. 9a). The SNR data from 2 November show that there are thermals present during the afternoon hours, similar to the summer cases. The development of the stable layer can also be seen, but it is broad and shallow, reflective of the initially shallow nocturnal boundary layer.
As in Fig. 3, but for 2 Nov 2008.
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
As in Fig. 3, but for 2 Nov 2008.
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
As in Fig. 3, but for 2 Nov 2008.
Citation: Journal of Applied Meteorology and Climatology 51, 3; 10.1175/2011JAMC2607.1
The sodar vertical velocities sometimes show the presence of the upward motion early in the time period, but this is not true for all cases. Without the detection of thermals, the use of the sodar to examine the ABL transition becomes less practical. For cases in which sodar data are available earlier in the day (e.g., 2 November; Fig. 9b), the vertical velocities indicate thermals in a similar way to the summer cases. The sodar horizontal wind profile is also less useful when the sodar is not running continuously. The winds for the cases included in this study were typically light and variable, with no dominant direction throughout the lowest 500 m of the atmosphere.
The overall pattern of the autumn ABL transition is similar to that of the summer ABL transition but occurs over a shorter time period (~1 h), with more overlap between parameters. The same surface variables can be examined to find at what time the transition is occurring and to study the characteristics before and after the transition (Table 3). The sodar becomes less useful during the autumn transition time period, however, because the more-subtle nature of the development of the NBL is less visible in the data for most cases.
4. Discussion
The characteristics of the ABL transition examined in this study can be compared with the results from previous studies to determine their validity in the Huntsville area. Because of a lack of direct surface heat flux measurements in our study, another definition of the AET was developed. In this case, one element of the transition was a decrease in temperature. This element corresponds to the definition used by Edwards et al. (2006). If the shortwave heating decreases, then temperature will decrease under clear skies if temperature advection is neutral. Because this situation occurs well before the formation of a temperature inversion, it extends the transition to an earlier time, as the Edwards et al. (2006) definition did.
Mahrt (1981) found that the low-level winds began to decelerate about 2 h before sunset. The measurements in that study were taken at a higher level (50–500 m) than the measurements in the current study. The same changes were found to occur, however. The 10-m wind speeds began to decrease at least 30 min before sunset. Unlike in Mahrt (1981), the conclusions from this study show no preference in the rotation of the wind as the nocturnal boundary layer develops.
The results from the sodar and 915-MHz profiler can be compared with the results found in Grimsdell and Angevine (2002) and Angevine (2008). The definition used in the first paper does not correspond to the results from the current study. The convective boundary layer height did not visibly decrease in the time–height series. The updated definition in Angevine (2008) also does not appear to correspond to the results of the current study. There was certainly a decrease in turbulence during the transition, but it was most visible within the convective boundary layer and was not limited to the top of the layer, as the definition states.
The combined use of observational data from a variety of instruments in this study shows that the early-evening boundary layer transition is made up of changes in several different atmospheric variables. Past studies have identified a subset of these changes, but this study has shown that a multisensor approach presents a more complete view of the transition. The use of both profiling and surface instruments provides high-temporal-resolution data throughout the depth of the boundary layer. This approach provides more information about the boundary layer transition than has been previously studied.
5. Conclusions
The AET was found to have a relatively consistent pattern, regardless of season. The AET often starts with a decrease in wind speed variance (which can be used to approximate TKE), followed by a temperature decrease, a mixing ratio increase, and a 10-m wind speed decrease, with increases in the wind speeds aloft. The variances are different for summer and autumn for each of the surface measurements other than wind, and therefore only the wind speed variances are used in the description of the boundary layer transition. The final stage of the AET is typically represented by a minimum in the sodar return and a coincident development of a stable stratified layer that is visible as a low-level maximum in sodar return. On the basis of this dataset, the AET duration over the entire boundary layer depth is about 160 min for the summer cases and is 85 min for the autumn cases. It has been shown in this study that the use of sodar and 915-MHz profiler backscatter and vertical velocity data can provide information about the AET from the convective boundary layer to the stable boundary layer that would not otherwise be available from only surface observations.
The definition of the AET developed in this paper should be considered as valid for cases with low 10-m wind speed and scattered cumulus clouds or clear conditions. We expect a longer and less-defined AET for cases of stronger wind (>5 m s−1) and persistence of clouds before sunset. Further research is necessary to determine how these characteristics might change if heavy cloud cover is present or under other atmospheric conditions.
Acknowledgments
This research has been funded by National Science Foundation Grant ATM-533596 and by the National Oceanic and Atmospheric Administration under Grant NA08OAR4600896.
REFERENCES
Acevedo, O. C., and D. R. Fitzjarrald, 2001: The early evening surface-layer transition: Temporal and spatial variability. J. Atmos. Sci., 58, 2650–2667.
Angevine, W. M., 2008: Transitional, entraining, cloudy, and coastal boundary layers. Acta Geophys., 56, 2–20.
Beare, R. J., J. M. Edwards, and A. J. Lapworth, 2006: Simulation of the observed evening transition and nocturnal boundary layers: Large-eddy simulation. Quart. J. Roy. Meteor. Soc., 132, 81–99.
Caughey, S. J., J. C. Wyngaard, and J. C. Kaimal, 1979: Turbulence in the evolving stable boundary layer. J. Atmos. Sci., 36, 1041–1052.
Cione, J. J., P. G. Black, and S. H. Houston, 2000: Surface observations in the hurricane environment. Mon. Wea. Rev., 128, 1550–1561.
Edwards, J. M., R. J. Beare, and A. J. Lapworth, 2006: Simulation of the observed evening transition and nocturnal boundary layers: Single-column modeling. Quart. J. Roy. Meteor. Soc., 132, 61–80.
Grant, A. L. M., 1997: An observational study of the evening transition boundary-layer. Quart. J. Roy. Meteor. Soc., 123, 657–677.
Grimsdell, A. W., and W. M. Angevine, 2002: Observations of the afternoon transition of the convective boundary layer. J. Appl. Meteor., 41, 3–11.
Karan, H., 2007: Thermodynamic and kinematic characteristics of low-level convergent zones observed by the Mobile Integrated Profiling System. Ph.D. dissertation, University of Alabama in Huntsville, 185 pp.
Mahrt, L., 1981: The early evening boundary layer transition. Quart. J. Roy. Meteor. Soc., 107, 329–343.
Nieuwstadt, F. T. M., and R. A. Brost, 1986: The decay of convective turbulence. J. Atmos. Sci., 43, 532–546.
Stull, R. B., 1988: An Introduction to Boundary Layer Meteorology. Kluwer Academic, 666 pp.
