The planetary boundary layer (PBL) reaches a height varying from less than 100 m to several kilometers above Earth’s surface and responds to surface forcing on time scales of about 1 h or less (Stull 1988). An important parameter that measures the depth of this layer and plays a key role in the mixing that takes place within the PBL is the PBL height. The PBL height depends on net radiation, land surface characteristics, and atmospheric moisture, among other quantities. The PBL height is important for many applications as detailed in the NASA PBL Incubation Study Team Report (Teixeira et al. 2021). An example of an important application is air quality, as the PBL height largely determines the concentration and distribution of constituents in the air, making it relevant to human health (Pérez et al. 2020; Haman et al. 2014; Miao et al. 2019). Other applications include wildfire monitoring (Potter 2012; Strada et al. 2012), agriculture (Wulfmeyer et al. 2015), weather prediction (Cohen et al. 2017; Mukhopadhyay et al. 2020) and climate, prediction of sea states (James et al. 2018), transportation (Watkins et al. 2010), hydrological modeling (Mockler et al. 2016; Santanello et al. 2017), and wind and solar energy (James et al. 2018; Emeis 2014). These applications all highlight the importance of accurate observational estimates of PBL height. Teixeira et al. (2021) made clear that the advancement of PBL science and applications requires a set of comprehensive global measurements. A global dataset allows for wider-scope geographical analysis of hourly PBL heights and provides the basis for a complete study of the connections of PBL height to seasonality, to latitude, proximity to oceans, and synoptic variability, and hence contributes to the advancement of PBL science and applications.
In this manuscript we describe a new surface-based PBL height dataset called Global Radar Wind Profiler Planetary Boundary Layer Height (GRWP-PBLH). PBL heights were estimated using existing data obtained from radar wind profiler (RWP) measurements at a large number of stations distributed all over the world. Molod et al. (2015, 2019) developed an algorithm that estimates PBL height using backscatter data measured by RWPs and applied it to data from the NOAA Profiling Network stations over the central United States. The dataset presented here builds on that work by expanding the geographical scope. To this end, we used a dataset compiled by the Met Office (UKMO) and published by the National Centre for Atmospheric Science British Atmospheric Data Centre (NCAS BADC). The geographical and temporal scope of this dataset provides a unique opportunity for producing the most globally and temporally expansive set of PBL height estimates to date. These data and the software to process them are now available to the public, following principles of open science.
The present manuscript serves to detail the data processing and production of GRWP-PBLH, the auxiliary signal-to-noise ratio (GRWP-SNR) data, and the software to aid in the production and use of these data. Following this introduction, the second section provides a summary of the use of RWP data to estimate PBL heights and describes the algorithm used in the present study. The third section includes a description of the raw data and their preprocessing. The fourth section details the GRWP-SNR data, the process of retrieving PBL height estimates, and the production of the GRWP-PBLH data. The fifth section presents a list and a description of the software developed for data visualization, as well as illustrative visual examples of these visualizations. A brief summary of this manuscript, data products, and accessibility of the GRWP-PBLH data and software are presented in the sixth section.
Radar wind profiler estimates of PBL height
The PBL height is a geophysical quantity whose definition and estimation are not as straightforward as they are for quantities such as temperature and water vapor. Many different methods exist to estimate PBL height based on measurements from many different instruments, each with their own advantages and disadvantages. Space-based measurements, such as those from the Cloud–Aerosol Lidar with Orthogonal Polarization (CALIOP) instrument aboard the Cloud–Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO) satellite (Jordan et al. 2010; McGrath-Spangler and Denning 2012, 2013; Luo et al. 2014, 2016), and those from the Global Navigation Satellite System–Radio Occultation (GNSS-RO) techniques (von Engeln et al. 2005; Ao et al. 2012) are global in spatial coverage but are limited in temporal resolution. Suborbital and surface-based PBL height retrievals have been obtained from measurements using radiosondes (Seidel et al. 2012), ceilometers (Caicedo et al. 2020), lidars (Lewis et al. 2013), radar wind profilers (Molod et al. 2015), microwave radiometers (de Arruda Moreira et al. 2019), Atmospheric Emitted Radiance Interferometer (AERI; see, for example, Turner and Löhnert 2014), and from combinations of instruments such as in the work reported in Duncan et al. (2022). These suborbital and surface-based observations are generally highly localized, but provide high temporal resolution. The importance and usefulness of surface-based observations that can be utilized to estimate PBL height is well established and is reviewed in detail by Kotthaus et al. (2023).
RWPs are particularly useful for PBL height estimation because although they provide localized measurements, there are extensive networks worldwide and so provide data to estimate PBL heights that are global in scope. The RWP is an active remote sensing instrument that can routinely observe wind and turbulence in the troposphere through scattering from clear-air irregularities of the atmospheric refractive index (Balsley and Gage 1980). RWPs emit a radar signal and measure the return signal or backscatter every 50–250 m throughout the vertical column of air. Most instruments compute and archive a signal-to-noise ratio (SNR). The RWP backscatter or SNR is sensitive to gradients in air density, and the discontinuities that mark the PBL height appear as maxima in the SNR. RWPs are limited by the lowest height at which measurements are taken, normally around 200 m. In addition, the instrument can also be sensitive to particles, hydrometeors, and insects (Ecklund et al. 1999). Many studies using different retrieval methods have shown that RWPs can provide reliable PBL height estimates (Angevine et al. 1994; Lemone et al. 1999; Angevine et al. 2001; Bianco and Wilczak 2002; Heo et al. 2003; Lee and Kawai 2011; Molod et al. 2015) at a relatively fine vertical resolution in the data. These retrieval methods use backscatter or SNR data from wind profilers to estimate PBL heights by identifying maxima in the backscatter or SNR profiles that are associated with changes in gradients to locate the top of the boundary layer.
Molod et al. (2015) developed a simple algorithm to estimate planetary boundary layer height using a long record of backscatter data available from the NOAA wind profiler network over the U.S. Central Plains region. The underlying assumption for their algorithm, as well as for most lidar or RWP PBL height algorithms, is that the gradients of moisture, hydrometeors, or particles at or near the PBL height create density gradients, and so Bragg scatter (in the case of RWPs only), that are manifest as maxima in the signal backscatter at the detector. Two unique elements of Molod et al. (2015) algorithm are 1) the establishment of the “emergence time” of the PBL height defined as the time of day at which the PBL rises from its nocturnal value into the range of the RWP detection and 2) the method to establish which among many potential signal maxima in a vertical profile represents the PBL height, referred to as the “true maximum.” Molod et al. (2019) modified the earliest version of the algorithm including adjustments to both of these elements as well as a stricter determination of when PBL heights are successfully retrieved. The study used the hourly clear-sky PBL height estimates based on the 20-yr period (1992–2012) of NOAA’s wind profiler network data to examine the behavior and variability of PBL heights over the United States. These studies showed that the algorithm is robust, relies on the existence of publicly available data, is relatively simple and not site specific.
The algorithm developed, detailed, and modified by Molod et al. (2015, 2019) underwent further adjustments summarized below. These adjustment were needed to accommodate the incorporation of measurements made by RWPs of widely varying characteristics (frequency and power, for example) that result in varied vertical resolution and temporal frequency of reporting, and to accommodate the change from backscatter to SNR data. The first major adjustment was to include the new step of interpolation of the raw data as detailed in the “RWP data source and its preprocessing” section. Hereafter, we refer to this latter version of the algorithm as the GRWP algorithm. The GRWP algorithm follows the approach previously established to first determine the “emergence time” and then examine the profile of SNR at each time to identify the height of the “true maximum” in the profile to be assigned to the PBL height at that time. We examined the sensitivity to the various tunable parameters and chose the values reported here.
The “emergence time” is defined in the algorithm as the time at which the SNR value at the lowest range gate is at its daily maximum. This definition is meant to capture the time at which the PBL height, represented by a maximum in SNR, is at the level of the bottom range gate. To determine the lowest range gate where the maximum SNR occurs, we search for the lowest level in the interpolated SNR data (matrix) that has at least 35% of its values defined, and called this level the emergence height. If this condition is not met in the bottom four levels, we do not define an “emergence time” and all PBL heights of that day are set to undefined. The time series at the emergence height is then smoothed with a moving average that attenuates approximately 25% of energy of the input data. The time at which the maximum SNR value in the smoothed emergence height time series is reached is then set as the “emergence time” for that day.
The search for the PBL height as the maximum in the vertical profile of SNR signal proceeds after the “emergence time” until 2000 local time (LT). If only one maximum is identified in the vertical profile, its height is set as the height of the PBL. For most vertical profiles, however, multiple local maxima exist and a criterion to reject small variations in the SNR data in favor of a “true maximum” is needed. To do so, the absolute maximum in the vertical profile is first identified. If there are undefined values (that signify missing data) below the absolute maximum in the profile, the PBL height is set to undefined. Starting with the lowest local maximum in the profile, if the difference between the local maximum and local minima above and below is greater than the standard deviation of the profile up to the local maximum itself, then the level of the maximum is selected as the height of the PBL and no further searching is necessary. If this condition is not satisfied, then the process of evaluating local maxima proceeds upward in the column until a “true maximum” is found and the PBL height assigned to that maximum. If this test fails for all local maxima identified in a profile, the PBL height is undefined for that time of day.
RWP data source and its preprocessing
RWP data source.
The present work utilizes SNR data from RWPs compiled and formatted by the UKMO and published by the NCAS BADC (UKMO 2008). These data can be accessed from https://catalogue.ceda.ac.uk/uuid/9e22544a66ba7aa902ae431b1ed609d6, upon creating an account and submitting a request. More information about the observations from wind profiling systems provided by UKMO (frequency of observations, list of parameters, type of stations) can be found at https://artefacts.ceda.ac.uk/badc_datadocs/ukmo-metdb/winpro.html. Approximately 400 wind profiler stations operated and maintained by individual national meteorological organizations throughout the world report their data to UKMO. All stations report wind data but many do not report SNR data and will not be included here since our algorithm requires SNR data to estimate PBL heights.
The BADC reports SNR data from 282 stations throughout the 26 countries of Australia, Austria, Belgium, Canada, Croatia, Czechia, Finland, France, Germany, Hungary, Ireland, Italy, Japan, the Netherlands, Norway, Poland, Portugal, Reunion, Samoa, Slovenia, Spain, Sweden, Switzerland, the United Kingdom, and the United States. The RWPs operating at these stations vary in make, model and operating frequencies, and the reported measurements vary in hours of reporting, frequency of measurement, and heights at which measurements are collected. Detailed information about the networks reporting data to the BADC, including information about the instruments used and modes of operations, can be found in Nash and Oakley (2001) (Europe), Dolman et al. (2018) (Australia), and Ishihara et al. (2006) (Japan). BADC has been publishing these data since 17 May 2009 and continues through the present. After the extensive quality control of these data detailed in the following subsection, the compiled dataset presented here includes a subset of these stations only. A map with the locations of the stations used in the present work is shown in Fig. 1. The present algorithm (of Molod et al. 2015) was not successful in retrieving PBL heights from valid SNR data reported at several stations (indicated by red open circles in the figure). These SNR data are included in the GRWP database for potential use by an alternative algorithm.
Map showing the locations of stations with data used in the GRWP-PBLH dataset. Stations marked with open blue circles correspond to locations where the GRWP algorithm successfully estimated PBL heights, and stations marked with open red circles correspond to locations where the algorithm failed to estimate PBL heights.
Citation: Bulletin of the American Meteorological Society 104, 5; 10.1175/BAMS-D-22-0002.1
The UKMO raw data are organized in directories by year, which contain subdirectories for each month. The month subdirectories contain one comma-separated value (CSV) file for each day. Each CSV file contains a header for each station with information about the reporting station followed by the values of the measured variables at varying heights (constituting a single profile) from that station. These variables are the time stamp, averaging time period, pressure, temperature, precipitation, humidity, date of record creation (history), height, wind profiler mode information, wind profiler quality control (QC) information, wind u component and QC information, wind υ component and QC information, wind w component and QC information, horizontal wind speed standard deviation, vertical wind speed standard deviation, and SNR. Information about the station header and the profile variables recorded in the raw data can be found at https://artefacts.ceda.ac.uk/badc_datadocs/ukmo-metdb/winpro.html.
Data preprocessing for PBLH retrieval.
While the raw data files have all the necessary information to retrieve PBL heights, the structure is quite cumbersome to use for the retrieval calculations. We therefore restructured the data to ease the computational burden of the PBL height retrievals. The SNR data files underwent further preprocessing involving incorporating cloud fraction data, selecting pertinent SNR data, interpolating the data, and reformatting from CSV to Cooperative Ocean/Atmosphere Research Data Service (COARDS)-compliant netCDF format for the creation of the GRWP-SNR files available to the public. These additional preprocessing steps increased the efficiency and user-friendliness of the GRWP-PBLH software and data. Figure 2 shows a flow diagram of the steps from preprocessing to the production of the GRWP-PBLH dataset. Details of this procedure are described below.
Flow diagram depicting the steps followed in the preprocessing and production of the GRWP-PBLH dataset. Details of these steps are in the “RWP data source and its preprocessing” and “GRWP datasets” sections.
Citation: Bulletin of the American Meteorological Society 104, 5; 10.1175/BAMS-D-22-0002.1
The raw data were restructured from files that contain data from all the stations for a single day into files that contain data from a single station for an entire year. Restructuring the raw data into one CSV file per station per year is done using a Python script readUKMOraw.py, which reads through the raw data files for an entire year and collects the profiles into a new file for each station. These restructured files contain a first row with column headings and subsequent rows containing the UTC time stamp, averaging time period, surface pressure, surface temperature, precipitation, humidity, history, height, wind profiler mode information, wind profiler QC information, wind u component and QC information, wind υ component and QC information, wind w component and QC information, horizontal wind speed standard deviation, vertical wind speed standard deviation, and SNR.
Some of the resulting restructured files had only “undefined” SNR values. We identified these files and removed them from the dataset, using the Python script separateNAdata.py, which copies only the files with useful data into a new directory. Finally, because the data are reported in UTC time but PBL height analyses often rely on local time, the Python script addLocalTime.py calculates the local time and adds a new column to the restructured files.
The remaining preprocessing steps were performed using the MATLAB script createGRWPsnr.m utilizing restructured CSV files. Due to the dependence of PBL height on cloud cover in general (Sedlar et al. 2022), information about the cloud fraction at each time of SNR measurement, when available, is provided in our dataset. We include the Clouds and the Earth’s Radiant Energy System (CERES) 3-hourly cloud fraction data available at https://ceres-tool.larc.nasa.gov/ord-tool/jsp/SYN1degEd41Selection.jsp as a general guidance for the net surface radiation at individual RWP sites. The GRWP-PBLH software script getcleardates.m is called by createGRWPsnr.m and loads the CERES data when available to obtain the total cloud fraction including low (surface–700 hPa), mid–low (700–500 hPa), mid–high (500–300 hPa), and high (300 hPa–tropopause) cloud fractions for each 3-h period for each station. The same cloud fraction is assigned to each SNR profile during the CERES 3-h window. The resolution of the CERES dataset is 1° × 1° global grid. The cloud fraction data at a given station correspond to the information provided by CERES for the grid box that contains the station. If cloud-cover data are not available, the cloud fraction values are set to undefined. At the time of writing of this manuscript, the CERES data were only available through July 2020.
The height of the lowest range gate of most RWPs limits their ability to provide data that can be used to estimate the PBL height of nocturnal shallow layers. Because of this limitation, we selected daytime SNR measurements between 0900 and 2000 local time, accounting for daylight standard time (DST), from the input restructured CSV files. During this time range the PBL height may be high enough to be detected at the lowest range gate. The selected data were further limited to measurements below 4,000 m, above which we do not expect to detect PBL heights. The final preprocessing step of vertically interpolating the SNR data before writing the data to GRWP-SNR netCDF files is also included as part of createGRWPsnr.m. The details of the interpolation are included in the GRWP-SNR dataset description in the “GRWP-SNR” section below.
GRWP datasets
The GRWP-PBLH data products include the GRWP-SNR netCDF and the GRWP-PBLH netCDF files. As described in the “RWP data source and its preprocessing” section and indicated in the diagram of Fig. 2, the GRWP-SNR files are produced using the restructured raw data (CSV format) and cloud fraction data from CERES. The GRWP-PBLH files are produced from the GRWP-SNR netCDF files (or from the restructured CSV files) using GRWP-PBLH software to obtain PBL height retrievals. The final datasets have the geographical scope shown in Fig. 1. Information about the specific temporal scope of the SNR and PBL height data available for each individual station included in the final datasets can be found in the GRWP-PBLH readme.pdf file, on pages 7, 8, 13, and 14, available at https://grwp-pblh.hunter.cuny.edu/. Overall, the GRWP datasets cover a period from 17 May 2009 through July 2020 although not all stations have reported data for the entire period.
GRWP-SNR.
The raw SNR data are reported at irregular time and height intervals. The PBL height retrieval algorithm used in the study relies on vertical gradients of SNR and so requires SNR data at regular height intervals. Therefore, the SNR data provided in the GRWP-SNR netCDF files have been linearly interpolated to obtain regular spacing in height. This vertical interpolation is done by the MATLAB script createGRWPsnr.m. To begin the interpolation process, the SNR data are placed into the appropriate location in a time–height grid at a resolution of 1 m, based on the observed height and time. This grid is then populated with values interpolated and extrapolated linearly in height. This process may result in extrapolations far beyond observed values in areas where there were missing measurements. For example, if one profile had measurements from 200 to 3,500 m and the next profile had measurements from 500 to 3,000 m, the values in the grid at the time of the latter profile between 200 and 500 m and between 3,000 and 3,500 m would be extrapolated by over 300 m. To prevent “overextrapolation,” any data that were extrapolated more than 50 m below or above a point of reported measurement were set to undefined. The available GRWP-PBLH software detailed in the companion readme.pdf file has also the capability to interpolate the GRWP-SNR data in time to a 1-min resolution to aid in the visualization of the SNR measurements.
The GRWP-SNR netCDF files contain latitude, longitude, height, local time, UTC time, SNR values, a mask to identify the heights at which SNR measurements were originally reported, and total cloud fraction. The mask to identify grid points of the originally reported SNR values is a grid populated with zeros and ones. A multiplication of each value of SNR by the value of the mask will return a grid populated with SNR values only at the heights and times of recorded measurements and zeros in the rest of the grid points. Each new netCDF file contains data for one year at one station and is named using the convention GRWP_SNRgrid_stationNumber_year.nc. The complete GRWP-SNR dataset includes 901 netCDF files from 106 stations. The GRWP-SNR netCDF files and the GRWP-PBLH software used to produce them are available at https://grwp-pblh.hunter.cuny.edu/ in the directories “GRWP_SNR_NetCDFs” and “GRWP_Software,” respectively.
GRWP-PBLH.
The GRWP algorithm, described in the “Radar wind profiler estimates of PBL height” section, is applied to the interpolated profiles (input) provided in the GRWP-SNR netCDF files. One of the GRWP-PBLH software files, createGRWPpblh.m, controls this entire process. createGRWPpblh.m first loads the GRWP-SNR netCDF files for all years for a specified station. The GRWP algorithm is then applied one day at a time to each SNR profile within that day, after the emergence time and before 2000 local time, and a PBL height estimate is retrieved. Once an entire day’s PBL heights are retrieved, each diurnal cycle is smoothed using a third-degree polynomial fit. The smoothing is necessitated by occasional profiles for which the algorithm chooses a “true maximum” that is not the PBL height. Extensive sensitivity tests were performed with higher-order polynomial interpolation, but we chose the third order so as to minimize the creation of spurious maxima while retaining the PBL height growth rate. The PBL heights reported in the GRWP-PBL dataset are sampled hourly from the third-order polynomial. The raw PBL height retrieval, before smoothing, can be easily obtained and visualized using the software provided with the dataset, following the instructions in the accompanying readme.pdf file. Additionally, the cloud fractions loaded from the GRWP-SNR netCDF files are added to allow for analysis of PBL height during varying cloud conditions.
As mentioned earlier, the GRWP algorithm uses vertical gradients of SNR to retrieve PBL height. Therefore, estimates of the error associated with these retrievals were set as twice the hourly vertical resolution of the SNR data. Since the vertical resolution of the recorded SNR data varies (within a profile, within a single day, and/or at a single station), we obtained the average vertical resolution of each profile, built diurnal cycles of vertical resolution, smoothed these diurnal cycles of vertical resolution using a third-degree polynomial, and sampled them hourly to obtain a vertical resolution estimate associated with each PBL height estimate.
We selected to provide the negative-SNR-value indicator because the original algorithm was developed to handle positive backscatter measurements and we believe that PBL height estimates obtained from profiles containing negative SNR values should be treated with caution. The negative-SNR-value indicator was set to a value of “1” if there were one or more negative SNR values recorded at altitudes below the absolute maximum in the SNR signal for the profile. Because the reported profiles do not necessarily occur hourly, the negative-SNR-value indicator nearest in time to the PBL height estimate was reported.
The hourly PBL height estimates (meters above ground level) were saved to a netCDF file along with latitude, longitude, altitude of the station, UTC time, local time, cloud fraction, error estimate, and the negative-SNR-value indicator. These netCDF files have a naming convention of GRWPPBLH_stationNum.nc. The complete GRWP-PBLH dataset includes 91 netCDF files from 91 stations. The GRWP-PBLH netCDF files and the GRWP-PBLH software used to produce them are available at https://grwp-pblh.hunter.cuny.edu/ in the directories “GRWP_PBLH_NetCDFs” and “GRWP_Software,” respectively.
Visualizations
The GRWP-SNR netCDF and GRWP-PBLH netCDF files can easily be used to produce figures to aid basic analyses of PBL heights data. This section lists and briefly describes the available GRWP-PBLH software for visualizations of SNR profiles, SNR contours, PBL height time series, and PBL height mean fields. The names of all scripts that produce visualizations begin with the word “main” and can be accessed at https://grwp-pblh.hunter.cuny.edu/ in the directory “GRWP_Software.” The published readme.pdf file contains further details on how to use the GRWP software.
SNR profiles and contours.
The GRWP-PBLH software script main_plotProfilesContours.m produces a figure with two panels, showing individual profiles and the diurnal cycle for a selected day of data. The panel on the left contains selected SNR profiles throughout the day with PBL height retrievals highlighted on each profile. The panel on the right contains the corresponding contoured SNR data with an overlay of the retrieved PBL height values for the selected day. Before running the script that produces this figure, the GRWP-PBLH software script loadSNRdata.m must be run with a user-defined option to load the selected day’s data from either the restructured CSV or the GRWP-SNR netCDF file into memory. The script main_plotProfilesContours.m also includes a user-defined option to either apply the GRWP algorithm to SNR data or to read the PBL height estimates from the published GRWP-PBLH netCDF files. Choosing to retrieve the PBL height estimates allows the resulting figure to include profile-by-profile PBL height retrievals as opposed to only the published smoothed hourly PBL height retrievals. Figure 3 shows an example output from this script for station 10394 (52.21°N, 14.13°E) located in Germany on 16 July 2018. The command to produce this figure utilized the options to load the SNR data from the GRWP-SNR netCDF file and to retrieve smoothed and unsmoothed PBL height estimates by running the algorithm as opposed to reading the smoothed PBL height estimates from the GRWP-PBLH netCDF file.
Example of diurnal evolution of PBL height on 16 Jul 2018 at station 10394 (52.21°N, 14.13°E, Germany). (a) Selected RWP SNR profiles (dB) throughout the day at the times indicated in the legend. Each successive profile has a +10-dB offset for visual clarity. Open circles on each profile indicate the PBL height estimated by the RWP algorithm. (b) Shading indicates SNR strength, the black curve with open circles corresponds to the PBL heights retrieved by the RWP algorithm, and the white line corresponds to the third-degree polynomial smoothing of that curve. “ET” denotes emergence time and “EH” denotes emergence height.
Citation: Bulletin of the American Meteorological Society 104, 5; 10.1175/BAMS-D-22-0002.1
PBL height time series.
The GRWP-PBLH software script main_plotPBLHTimeSeries.m produces a figure containing all PBL height diurnal cycles from the GRWP-PBLH netCDF file for a selected station and a selected year. The user may zoom into a desired time period to more easily view a subset of the time series. The diurnal cycles are colored to represent the cloud-cover conditions associated with the retrievals. Clear skies are defined here as having a total cloud fraction of less than 0.15, cloudy skies as having a total cloud fraction greater than 0.85 and partially cloudy skies include all values of cloud fraction in between 0.15 and 0.85. This script provides an option to include or exclude the PBL height retrievals associated with SNR profiles containing negative values below the absolute maximum in their signals. Figure 4 shows an example output from this script for station 47674 (35.15°N, 140.31°E, Japan) for the year 2010 including only retrievals associated with positive SNR values. It has been zoomed in to show only the month of July.
Example of a discontinuous time series of PBL heights (m) for station 47674 (35.15°N, 140.31°E, Japan) produced by main_plotPBLHTimeSeries.m, zooming in to show the portion of the series from 1 to 31 Jul 2010. Missing times indicate that there were no SNR data for that time or that the GRWP algorithm was unable to retrieve PBL heights from the SNR profiles of that time.
Citation: Bulletin of the American Meteorological Society 104, 5; 10.1175/BAMS-D-22-0002.1
Mean fields.
The scripts main_plotMeans.m, main_plotSeasonalMeansComparison.m, and main_plotCloudMeansComparison.m published in the GRWP-PBLH software produce figures of hourly PBL height mean fields calculated from GRWP-PBLH netCDF files. The figures also display the error estimate of the hourly PBL height means, which is twice the mean of the vertical resolutions associated with each included PBL height retrieval. These scripts, described below, allow the user to specify the desired data to be included in calculating the mean field. In addition, the scripts allow for the graphic inclusion of the number of PBL height retrievals (count) that were used in the calculation of a mean field at each hour.
main_plotMeans.m allows the user to select the PBL height retrievals associated with a station, year(s), month(s), or season(s), cloud level, and negative-SNR indicator to include in the calculation of a mean.
main_plotSeasonalMeansComparison.m produces a plot of the mean hourly PBL height retrievals for winter, spring, summer, and fall or for January, April, July, and October. Climatological mean PBL heights are only provided when the sample size exceeds a threshold of 10% of the possible number of values. This script allows the user to select the PBL height retrievals associated with a station, year(s), months, or seasons, cloud level, and negative-SNR indicator to include in the mean calculation. Figure 5a shows an example output from this script (modified to remove the annotations—see readme.pdf) utilizing the options to include retrievals associated with all cloud levels and only positive SNR signal values from the characteristic months at station 10394 (52.21°N, 14.13°E, Germany) for the years 2009–20.
main_plotCloudMeansComparison.m produces a plot of the mean hourly PBL height retrievals associated with clear skies and with cloudy skies. This script allows the user to select the PBL height retrievals associated with a station, year(s), month(s), or season(s), and negative-SNR indicator to include in the mean calculation. Figure 5b shows an example output from this script utilizing the options to include retrievals associated only with positive SNR signal values using all data available for the period 2009–20 at station 10394 (52.21°N, 14.13°E, Germany).
Examples of mean diurnal cycles of PBL height retrieved using the GRWP algorithm and visualized with the GRWP-PBLH software described in the “GRWP datasets” section. Data are from station 10394 (52.21°N, 14.13°E, Germany). (a) Climatological mean diurnal cycles for the months of January, April, July, and October. Retrievals used to calculate these mean fields were obtained for all cloud-cover conditions and using only positive SNR signal values. We used 123 days of data to compute January mean, 151 days for the April mean, 179 for the July mean, and 161 days for the October mean. (b) Climatological means of retrieved PBL heights for clear and cloudy skies: cloud fraction of less than 0.15 and greater than 0.85, respectively. The clear-skies mean included retrievals from 214 days and the cloudy-skies mean from 230 days. In both panels the error estimate is equal to twice the average vertical resolution of the profiles used for PBL height retrievals at each hour.
Citation: Bulletin of the American Meteorological Society 104, 5; 10.1175/BAMS-D-22-0002.1
Summary
In this work we introduced GRWP-PBLH, a new dataset of PBL height estimates based on long-term measurements from networks of radar wind profilers distributed at many locations around the world. The algorithm of Molod et al. (2015, 2019) was successfully adjusted to be used with signal-to-noise ratio rather than backscatter data and to obtain PBL height estimates from irregularly sampled data. The complete GRWP-PBLH dataset is available to the community, with a supplementary dataset, GRWP-SNR, containing the corresponding SNR data from which the PBL heights were retrieved. A readme.pdf file and the software developed to process the data and perform useful calculations and visualizations are also available with the dataset. Table 1 summarizes the available GRWP-PBLH products.
Inventory and directory structure of GRWP-PBLH products.
The GRWP-PBLH dataset is the most comprehensive global-extent, long-term collection of surface-based PBL height data to date. PBL height retrievals are provided for the period May 2009–July 2020. There are, however, additional available RWP data that can be used to further expand the scope of the GRWP-PBLH dataset. For example, Molod et al. (2015) presented a set of PBL height estimates from the NOAA Profiling Network RWPs in the United States, which is not currently included in the GRWP-PBLH dataset. Additional data sources that may be incorporated in the GRWP-PBLH dataset in the future include the Atmospheric Radiation Measurement’s (ARM) RWP data and the Cooperative Agency Profilers’ (CAP) RWP data. The GRWP-PBLH dataset and software are made available to the public in keeping with open science standards at: https://grwp-pblh.hunter.cuny.edu/. The dataset will be maintained and updated by adding SNR and PBL height data each year as they become available from the raw data sources.
Acknowledgments.
The research for the present work was supported by NASA Grant 80NSSC20K0664. Partial support for H. J.’s and H. S.’s work was also provided by the PSC-CUNY Research Foundation Award (Enhanced) 64711-00. Author H. S. wishes to acknowledge the invaluable experience gained as a member of the NASA Incubation PBL Study Team and the collegiality of all members of the team. In addition, the authors are grateful to two anonymous reviewers for their careful scrutiny of the original manuscript and figures and for their constructive criticism, all of which led to revisions resulting in a much improved manuscript.
Data availability statement.
Raw data from the networks of radar wind profilers are available from the Met Office at https://data.ceda.ac.uk/badc/ukmo-metdb/data/winpro/. Cloud fraction data are available from NASA CERES at https://ceres-tool.larc.nasa.gov/ord-tool/jsp/SYN1degEd41Selection.jsp. The new GRWP-PBLH and GRWP-SNR datasets and companion software are available at https://grwp-pblh.hunter.cuny.edu/.
References
Angevine, W. M., A. B. White, and S. K. Avery, 1994: Boundary-layer depth and entrainment zone characterization with a boundary-layer profiler. Bound.-Layer Meteor., 68, 375–385, https://doi.org/10.1007/BF00706797.
Angevine, W. M., H. Baltink, and F. Bosveld, 2001: Observations of the morning transition of the convective boundary layer. Bound.-Layer Meteor., 101, 209–227, https://doi.org/10.1023/A:1019264716195.
Ao, C. O., D. E. Waliser, S. K. Chan, J.-L. Li, B. Tian, F. Xie, and A. J. Mannucci, 2012: Planetary boundary layer heights from GPS radio occultation refractivity and humidity profiles. J. Geophys. Res., 117, D16117, https://doi.org/10.1029/2012JD017598.
Balsley, B. B., and K. S. Gage, 1980: The MST radar technique: Potential for middle atmospheric studies. Pure Appl. Geophys., 118, 452–493, https://doi.org/10.1007/BF01586464.
Bianco, L., and J. M. Wilczak, 2002: Convective boundary layer depth: Improved measurement by Doppler radar wind profiler using fuzzy logic methods. J. Atmos. Oceanic Technol., 19, 1745–1758, https://doi.org/10.1175/1520-0426(2002)019<1745:CBLDIM>2.0.CO;2.
Caicedo, V., R. Delgado, R. Sakai, T. Knepp, D. Williams, K. Cavender, B. Lefer, and J. Szykman, 2020: An automated common algorithm for planetary boundary layer retrievals using aerosol lidars in support of the U.S. EPA Photochemical Assessment Monitoring Stations program. J. Atmos. Oceanic Technol., 37, 1847–1864, https://doi.org/10.1175/JTECH-D-20-0050.1.
Cohen, A. E., S. M. Cavallo, M. C. Coniglio, H. E. Brooks, and I. L. Jirak, 2017: Evaluation of multiple planetary boundary layer parameterization schemes in southeast U.S. cold season severe thunderstorm environments. Wea. Forecasting, 32, 1857–1884, https://doi.org/10.1175/WAF-D-16-0193.1.
de Arruda Moreira, G., and Coauthors, 2019: Analyzing the turbulent planetary boundary layer by remote sensing systems: The Doppler wind lidar, aerosol elastic lidar and microwave radiometer. Atmos. Chem. Phys., 19, 1263–1280, https://doi.org/10.5194/acp-19-1263-2019.
Dolman, B. K., I. M. Reid, and C. Tingwell, 2018: Stratospheric tropospheric wind profiling radars in the Australian network. Earth Planets Space, 70, 170, https://doi.org/10.1186/s40623-018-0944-z.
Duncan, J. B., and Coauthors, 2022: Evaluating convective planetary boundary layer height estimations resolved by both active and passive remote sensing instruments during the CHEESEHEAD19 field campaign. Atmos. Meas. Tech., 15, 2479–2502, https://doi.org/10.5194/amt-15-2479-2022.
Ecklund, W. L., C. R. Williams, P. E. Johnston, and K. S. Gage, 1999: A 3-GHz profiler for precipitating cloud studies. J. Atmos. Oceanic Technol., 16, 309–322, https://doi.org/10.1175/1520-0426(1999)016<0309:AGPFPC>2.0.CO;2.
Emeis, S., 2014: Current issues in wind energy meteorology. Meteor. Appl., 21, 803–819, https://doi.org/10.1002/met.1472.
Haman, C. L., E. Couzo, J. H. Flynn, W. Vizuete, B. Heffron, and B. L. Lefer, 2014: Relationship between boundary layer heights and growth rates with ground-level ozone in Houston, Texas. J. Geophys. Res. Atmos., 119, 6230–6245, https://doi.org/10.1002/2013JD020473.
Heo, B.-H., S. Jacoby-Koaly, K.-E. Kim, B. Campistron, B. Benech, and E.-S. Jung, 2003: Use of the Doppler spectral width to improve the estimation of the convective boundary layer height from UHF wind profiler observations. J. Atmos. Oceanic Technol., 20, 408–424, https://doi.org/10.1175/1520-0426(2003)020<0408:UOTDSW>2.0.CO;2.
Ishihara, M., Y. Kato, T. Abo, K. Kobayashi, and Y. Izumikawa, 2006: Characteristics and performance of the operational wind profiler network of the Japan Meteorological Agency. J. Meteor. Soc. Japan, 84, 1085–1096, https://doi.org/10.2151/jmsj.84.1085.
James, E. P., S. G. Benjamin, and M. Marquis, 2018: Offshore wind speed estimates from a high-resolution rapidly updating numerical weather prediction model forecast dataset. Wind Energy, 21, 264–284, https://doi.org/10.1002/we.2161.
Jordan, N. S., R. M. Hoff, and J. T. Bacmeister, 2010: Validation of Goddard Earth Observing System-version 5 MERRA planetary boundary layer heights using CALIPSO. J. Geophys. Res., 115, D24218, https://doi.org/10.1029/2009JD013777.
Kotthaus, S., and Coauthors, 2023: Atmospheric boundary layer height from ground-based remote sensing: A review of capabilities and limitations. Atmos. Meas. Tech., 16, 433–479, https://doi.org/10.5194/amt-16-433-2023.
Lee, S.-J., and H. Kawai, 2011: Mixing depth estimation from operational JMA and KMA wind-profiler data and its preliminary applications: Examples from four selected sites. J. Meteor. Soc. Japan, 89, 15–28, https://doi.org/10.2151/jmsj.2011-102.
Lemone, M. A., M. Zhou, C.-H. Moeng, D. H. Lenschow, L. J. Miller, and R. Grossman, 1999: An observational study of wind profiles in the baroclinic convective mixed layer. Bound.-Layer Meteor., 90, 47–82, https://doi.org/10.1023/A:1001703303697.
Lewis, J. R., E. J. Welton, A. M. Molod, and E. Joseph, 2013: Improved boundary layer depth retrievals from MPLNET. J. Geophys. Res. Atmos., 118,9870–9879, https://doi.org/10.1002/jgrd.50570.
Luo, T., R. Yuan, and Z. Wang, 2014: Lidar-based remote sensing of atmospheric boundary layer height over land and ocean. Atmos. Meas. Tech., 7, 173–182, https://doi.org/10.5194/amt-7-173-2014.
Luo, T., Z. Wang, D. Zhang, and B. Chen, 2016: Marine boundary layer structure as observed by A-train satellites. Atmos. Chem. Phys., 16, 5891–5903, https://doi.org/10.5194/acp-16-5891-2016.
McGrath-Spangler, E. L., and A. S. Denning, 2012: Estimates of North American summertime planetary boundary layer depths derived from space-borne lidar. J. Geophys. Res., 117, D15101, https://doi.org/10.1029/2012JD017615.
McGrath-Spangler, E. L., and A. S. Denning, 2013: Global seasonal variations of midday planetary boundary layer depth from CALIPSO space-borne lidar. J. Geophys. Res. Atmos., 118, 1226–1233, https://doi.org/10.1002/jgrd.50198.
Miao, Y., J. Li, S. Miao, H. Che, Y. Wang, X. Zhang, R. Zhu, and S. Liu, 2019: Interaction between planetary boundary layer and PM2.5 pollution in megacities in China: A review. Curr. Pollut. Rep., 5, 261–271, https://doi.org/10.1007/s40726-019-00124-5.
Mockler, E., K. Chun, G. Sapriza-Azuri, M. Bruen, and H. Wheater, 2016: Assessing the relative importance of parameter and forcing uncertainty and their interactions in conceptual hydrological model simulations. Adv. Water Resour., 97, 299–313, https://doi.org/10.1016/j.advwatres.2016.10.008.
Molod, A., H. Salmun, and M. Dempsey, 2015: Estimating planetary boundary layer heights from NOAA Profiler Network wind profiler data. J. Atmos. Oceanic Technol., 32, 1545–1561, https://doi.org/10.1175/JTECH-D-14-00155.1.
Molod, A., H. Salmun, and A. B. M. Collow, 2019: Annual cycle of planetary boundary layer heights estimated from wind profiler network data. J. Geophys. Res. Atmos., 124, 6207–6221, https://doi.org/10.1029/2018JD030102.
Mukhopadhyay, P., K. Raghavan, R. Nanjundiah, and M. Mohapatra, 2020: Prediction of extreme events: Current status and future pathways against the backdrop of climate change. Bull. Amer. Meteor. Soc., 101, E1137–E1141, https://doi.org/10.1175/BAMS-D-20-0037.1.
Nash, J., and T. Oakley, 2001: Development of COST 76 wind profiler network in Europe. Phys. Chem. Earth, 26, 193–199, https://doi.org/10.1016/S1464-1909(00)00239-2.
Pérez, I. A., A. Garcia, L. Sanchez, N. Pardo, and B. Fernandez Duque, 2020: Key points in air pollution meteorology. Int. J. Environ. Res. Public Health, 17, 8349, https://doi.org/10.3390/ijerph17228349.
Potter, B. E., 2012: Atmospheric interactions with wildland fire behaviour—I. Basic surface interactions, vertical profiles and synoptic structures. Int. J. Wildland Fire, 21, 779–801, https://doi.org/10.1071/WF11128.
Santanello, J., and Coauthors, 2017: Land–atmosphere interactions: The LoCo perspective. Bull. Amer. Meteor. Soc., 99, 1253–1272, https://doi.org/10.1175/BAMS-D-17-0001.1.
Sedlar, J., and Coauthors, 2022: Investigating the impacts of daytime boundary layer clouds on surface energy fluxes and boundary layer structure during CHEESEHEAD19. J. Geophys. Res. Atmos., 127, e2021JD036060, https://doi.org/10.1029/2021JD036060.
Seidel, D. J., Y. Zhang, A. Beljaars, J.-C. Golaz, A. R. Jacobson, and B. Medeiros, 2012: Climatology of the planetary boundary layer over the continental United States and Europe. J. Geophys. Res., 117, D17106, https://doi.org/10.1029/2012JD018143.
Strada, S., C. Mari, J.-B. Filippi, and F. Bosseur, 2012: Wildfire and the atmosphere: Modelling the chemical and dynamic interactions at the regional scale. Atmos. Environ., 51, 234–249, https://doi.org/10.1016/j.atmosenv.2012.01.023.
Stull, R. B., 1988: An Introduction to Boundary Layer Meteorology. Kluwer Academic, 666 pp.
Teixeira, J., and Coauthors, 2021: Toward a global planetary boundary layer observing system. NASA PBL Incubation Study Team Rep., 134 pp., https://science.nasa.gov/science-red/s3fs-public/atoms/files/NASA_PBL_Incubation_Final_Report_2.pdf.
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.
UKMO, 2008: Wind profiler observations, part of the Met Office MetDB system. NCAS British Atmospheric Data Centre, accessed 11 March 2023, https://catalogue.ceda.ac.uk/uuid/9e22544a66ba7aa902ae431b1ed609d6.
von Engeln, A., J. Teixeira, J. Wickert, and S. Buehler, 2005: Using CHAMP radio occultation data to determine the top altitude of the planetary boundary layer. Geophys. Res. Lett., 32, L06815, https://doi.org/10.1029/2004GL022168.
Watkins, S., M. Thompson, B. Loxton, and M. Abdulrahim, 2010: On low altitude flight through the atmospheric boundary layer. Int. J. Micro Air Veh., 2, 55–68, https://doi.org/10.1260/1756-8293.2.2.55.
Wulfmeyer, V., and Coauthors, 2015: A review of the remote sensing of lower tropospheric thermodynamic profiles and its indispensable role for the understanding and the simulation of water and energy cycles. Rev. Geophys., 53, 819–895, https://doi.org/10.1002/2014RG000476.
