1. Introduction
The West Texas Mesonet (WTM) is a joint partnership between the Atmospheric Science Group and Wind Science and Engineering Research Center at Texas Tech University. Originating in 1999, the main goal of the WTM is to provide free, timely, and accurate meteorological and agricultural data to all residents of the South Plains/Rolling Plains region of western Texas. The WTM has a close partnership with the National Weather Service (NWS) offices in Lubbock, Midland, Amarillo, and Fort Worth, and provides valuable real-time data to aid in warning verification for the NWS in the data-sparse region of western Texas.
The WTM, originally intended to be a pilot project for a statewide Texas Mesonet, currently (as of January 2004) consists of 40 automated surface meteorological stations, two atmospheric profilers, and one upper-air sounding system. All, except one, of the WTM surface stations are within 250 km of Lubbock, Texas (see Fig. 1). Each surface station measures up to 28 different parameters with an observation period of 5 min for meteorological data (e.g., air temperature, humidity, etc.) and 15 min for agricultural data (e.g., soil temperature, soil moisture content, etc.). All real-time data from the surface stations are available free of charge online at the WTM Web page at www.mesonet.ttu.edu. The WTM also includes two atmospheric profilers located at Reese Technology Center, Texas.
2. West Texas Mesonet surface observation systems
The WTM project was modeled after the Oklahoma Mesonet (Brock et al. 1995). The site selection specifications, site layout, and instrumentation selection of the WTM are nearly identical to those used by the Oklahoma Mesonet (Shafer et al. 1993). The original proposal for the WTM required the installation of at least one mesonet station per county in a 28-county region surrounding Lubbock, Texas. There are currently 40 stations in 28 counties as of January 2004. At least three additional WTM stations are planned for completion prior to the end of 2004.
a. Site selection
The WTM covers two distinct geographic areas in western Texas. A flat plateau region west of the Caprock Escarpment dominates the South Plains region that surrounds Lubbock. This area has a significant amount of irrigated agricultural production (mostly cotton) with a limited population outside of the main city of Lubbock. The land is very flat with minimal wind obstructions. The land east of the Caprock escarpment is called the Rolling Plains region, which is characterized by small hills and a very sparse population. This area is ranch country; again it contains only minimal wind obstructions.
There were three main constraints in determining site selection in the above regional area. The first constraint was to maintain a station spacing of approximately 35 km throughout the network while satisfying the physical constraints of a line of sight radio network. Stations were located at strategic points to act as both full mesonet stations and radio repeaters. This approach required two or three mesonet stations in one county while other counties have only one station. The WTM has a dense network of stations aligned north–south near the Caprock Escarpment to aid in radio reception between the stations on and off the Caprock. There are approximately 750 m in elevation change between the farthest northwest and southeast stations. Again this required some creative positioning of stations to aid in radio reception.
The second constraint in site selection was to maintain correct wind exposure and site slope. For each site, the Oklahoma Mesonet guidelines of maintaining a site slope of less than 5° and limiting obstruction heights to 0.05 times the distance between the obstruction and the anemometer were used (Brock et al. 1995). Occasionally this requirement has demanded the removal of mesquite and other small trees that violate the wind exposure guidelines from around the site.
The third constraint in site selection involved landowners providing the site free of charge. Approximately two-thirds of the WTM sites are on public land owned by either a local city or county. The remaining sites are on privately owned land. Landowners were asked to refrain from any major disruption in the surrounding land characteristics for a minimum of 10 yr. Permission was also obtained to access each site from each landowner in all weather conditions and at any time of the day.
b. Site layout and details
The layout of each mesonet station was modeled after the Oklahoma Mesonet with only minor differences in sensor positioning (see Figs. 2 and 3 for site schematics). Each station has a fenced-in 10 m by 10 m plot of land with a 10 m tall, guyed aluminum tower. The tower is hinged and can easily be lowered by two personnel for sensor replacement and maintenance. A 70 cm wide by 70 cm deep concrete base was used to secure the tower with three guy wires used for additional support.
Each station uses solar panels to charge several deep-cycle gel-type marine batteries for power. The number of solar panels and batteries at a site varies depending on the amount of radio repeater traffic at the individual location. An average site has two 20-Wsolar panels mounted with a south exposure and two deep-cycle marine batteries. Several major repeater sites have two 50-W solar panels and four deep-cycle marine batteries for power. To minimize shading on sensors, the solar panels are mounted approximately half way up the tower. No more than two 50-W solar panels are mounted on the tower because of wind loading concerns. If additional solar panels are needed, they are mounted at ground level approximately 6 m away from the tower itself.
The rain gauge (with a wind screen), soil temperature, and moisture content sensors are the only instruments not attached directly to the tower. The other sensors are mounted on boom arms at different levels on the tower (see Fig. 4). All wiring from the off-tower sensors is routed through PVC pipe and buried underground. The rain gauge is mounted to a 45 cm square concrete pad that is 20 cm above ground level to minimize flooding problems from heavy rainfall events. The natural and bare soil temperature/moisture content plots are marked with treated lumber to identify their location during mowing.
The majority of the WTM sites are located in rural areas, which at times are surrounded by cattle. Therefore, the fences had to be made of barbwire with stout fence posts. All fence posts are anchored into the ground with concrete, and additional dead-man braces are added for support. For those sites with no cattle, regular hog fencing was used with minimal fence posts. Tumbleweeds accumulate both inside and outside the fenced-in area. These tumbleweeds can accumulate to significant depths inside the fenced-in area during the spring months. At times, the tumbleweeds have blocked rain gauges and broken low-level sensors at several WTM locations. Tumbleweed removal is therefore part of the maintenance schedule for each site.
Animal damage to low-level sensors and wiring is a major problem. There is a significant concentration of black-tailed prairie dogs on the South Plains region of western Texas. Prairie dogs have chewed through any unprotected wiring near or below ground level. Additional protection has been provided to the low-level sensor wiring in the form of PVC conduit to minimize the damage. Predatory birds (hawks, owls, etc.) have also damaged sensors on boom arms mounted on the tower and destroyed wiring. Additional protection has been provided to sensor wiring and additional space has been allocated on various different boom arms for the birds to roost without damaging individual sensors.
Maintenance at each site is performed every two months during regularly scheduled visits. Regularly scheduled maintenance includes mowing the grass, removing debris, testing and calibrating instrumentation, and downloading data from the datalogger. The downloaded data are considered to be the official data archive, as data are occasionally lost during radio and phone transmissions.
c. Instrumentation and data acquisition equipment
Each WTM station has a standard set of meteorological sensors at identical positions on each tower. Soil and other agricultural sensors are not available at every site. See Table 1 for a complete list of sensors, and Table 2 for sensor characteristics, accuracies, and dataset resolutions.
1) Dataloggers
The WTM uses a Campbell Scientific CR23X datalogger with one megabyte of extended memory. Each datalogger records a 5-min observation from each of the meteorological sensors and a 15-min observation from each of the agricultural sensors. Sampling intervals vary from 3 to 60 s depending on the sensor. Data are stored in the logger in the 5- and 15-min observation format. Up to 78 days of data can be stored in the circular memory buffer of each logger before data are lost. Each datalogger has two 6-V backup batteries in case power is lost from the main deep-cycle marine batteries. This battery system is separate from the main batteries and can power the datalogger for approximately one month. These batteries help prevent the loss of data acquisition at each site.
2) Anemometers
The primary wind sensor employed at each WTM site is an R. M. Young 05103 Wind Monitor located at the 10-m height AGL (WMO 1983). This propeller-type anemometer measures both wind speed and direction. The information from this anemometer is used to provide each site’s average wind speed and direction and peak gust. Standard deviations of the wind speed and direction are also generated using the information from this anemometer. An additional anemometer was added to each site at the 2-m level to satisfy requests from the agricultural community for low-level wind data for crop spraying. The 2-m anemometer is an R.M. Young 03103 Wind Sentry. Only 5-min average wind speeds are recorded by the datalogger for this sensor. While the 2-m peak gust information would be beneficial for providing QA/QC checks on the data collected from the 10-m anemometer, it is currently not recorded because of radio transmission/power limitations.
3) Temperature and relative humidity probes
The primary temperature and relative humidity sensor is the Vaisala HMP45C (modified by Campbell Scientific) located at 1.5-m AGL (WMO 1983). The HMP45C is mounted in an unaspirated 12-plate radiation shield that is positioned 0.9 m away from the tower. This unit is very reliable and has long-term stability. Additionally, Campbell Scientific 107 Temperature (thermistor-type) probes were installed at both the 9- and 2-m level. These sensors are mounted in a nonaspirated 6-plate radiation shield that is positioned 0.6 m away from the tower. The use of nonaspirated radiation shields enables small radiation-induced errors to occur in the temperature measurements, especially in relatively calm conditions with low sun angles, high surface albedos, and high insolation. These errors have been documented in a number of publications including Brotzge and Crawford (2000) and Richardson (1995).
4) Solar radiation
Due to cost restraints, the WTM uses two types of pyranometers to measure incoming solar radiation. The Kipp and Zonen SP-Lite pyranometer is a third-class pyranometer used at the majority of stations. These units are inexpensive and provide reliable estimates of incoming solar radiation. The Kipp and Zonen CM-3 pyranometer is used on mesonet stations in a nine-county radius surrounding Lubbock. These are classified as a second-class pyranometer by the WMO (WMO 1983) and provide better estimates of incoming solar radiation. All mesonet pyranometers are returned to the manufacturer for calibration.
5) Rainfall measurements
The primary rain gauge in the WTM is the Hydrological Services TB3 siphon tipping-bucket rain gauge. This rain gauge has a 200-mm diameter receiver with a 0.0254-cm calibrated bucket to measure rainfall. A large portion of the gauge is made of aluminum, which has a higher survivability in hailstorms than the standard plastic tipping-bucket gauge.
The top of the TB3 gauge is located at the recommended height of 0.6-m AGL (Brock et al. 1995). A NovaLynx Alter style windscreen surrounds each TB3 rain gauge. The TB3 rain gauge is not heated and will not measure the correct amount of precipitation in snow or freezing rain events. Additional TB4 model rain gauges will be added to the WTM this year. The TB3 and TB4 model rain gauges are very similar with only minor changes in design. The TB4 does include a plastic base, but a special hail shield has been designed to protect the nonmetallic exterior.
6) Barometic pressure
The WTM uses a Vaisala PTB220 Class B digital barometer to measure station pressure at each site. The barometer is mounted inside the datalogger enclosure box at a height of 0.75-m AGL. Calculations of altimeter and sea level pressure are not generated at each site. The average station pressure over a 5-min period is currently the only pressure value recorded in the archive.
7) Soil temperature and moisture
The Campbell Scientific 107 (thermistor-type) probe is used to measure soil temperature values at different levels under natural and bare soil conditions. The natural soil plot has probes at 5, 10, and 20 cm below ground level. The bare soil plot has probes at 5 and 20 cm below ground level. The Campbell Scientific 615 Water Content Reflectometer is used to measure the volumetric water content at 5, 20, 60, and 75 cm below ground level. The reflectometer creates a square wave output that when combined with different calibration equations, based on soil type, density, and conductivity values at each depth, is used to estimate the water content. Soil samples for the various levels were given to a private laboratory that generated soil conductivity values. These values are then applied to the datalogger program at each site to generate correct readings. At times, adjustments are made following heavy rain events, which yield saturated soil conditions. Soil temperature and moisture data are not available for all WTM stations.
8) Leaf wetness sensor
The Campbell Scientific 237 Leaf Wetness sensor is used to provide an estimate of moisture on plants (Gillespie and Kidd 1978). The WTM positions the sensor at 50-cm AGL on the north-facing side of the tower. This was chosen to simulate droplet formation on mature cotton plants (shady side). Individuals from the agricultural community use these data to determine when to spray their crops. Data from the leaf wetness sensor can also be used to determine when dew and fog will develop.
d. Instrumentation calibration
Given funding limitations, the WTM only provides in-house calibration of two sensor types: anemometers and rain gauges. All other instrumentation is periodically returned to the manufacturer for calibration.
3. Boundary layer profilers
The Vaisala LAP-3000 radar profiler with RASS is used to obtain estimates of wind speed, wind direction, and virtual temperature. Data are collected at 30-min intervals using 60-m gate spacing from the surface to 3-km AGL. A REMTECH PA5 SODAR profiler is used to obtain estimates of wind direction, wind speed, vertical velocity, and additional static stability parameters. Data are collected at 30-min intervals using 50-m gate spacing from the surface to 1.5-km AGL. Archived profiler data are available upon request, while real-time data will be placed online in the near future.
4. Atmospheric sounding system
A Vaisala DigiCORA III upper-air sounding system is employed to launch meteorological soundings from Reese Center during interesting weather events. The DigiCORA III system records a variety of atmospheric parameters every 1.5 s. The system is flexible enough to allow both GPS and LORAN-ready radiosondes. Soundings are generally launched in response to requests from the NWS or in support of local and regional research projects. All data from the Reese Center launches are available online at the WTM Web page.
5. Communications
One goal of the WTM is to make 5-min observations available, in a real-time environment, to a wide variety of users. An approach similar to the Oklahoma Mesonet (Brock et al. 1995) was proposed to the State of Texas using the Texas Law Enforcement Telecommunications System (TLETS) for two-way communication between a data collection computer and remote mesonet stations. This proposal, however, was denied and other alternatives were pursued. Initially, recurring costs of any sort were not part of the WTM budget. Therefore, emphasis was placed on a two-way communications system that would be reliable and have a reasonable bandwidth, but not incur monthly fees. A summary of the current methods used for data transmission is provided in Fig. 5.
a. Extended Line of Sight Radio (ELOS)
Satellites were never seriously considered as a viable method for two-way communications. Generally slow data rates, cumbersome communication methods (Brock et al. 1995), and recurring expenses were unacceptable. ELOS communication provided a reasonable alternative with limited recurring expenses, but the availability of a high antenna site became a necessity. A 73-m tower was erected by the Wind Science and Engineering Research Center (Texas Tech University) at the Reese Technology Center. Two antennas, installed at the 61-m level, serve as the base radio station. A short-haul radio link is used to transfer all data to a data collection computer at the WTM facility approximately 3.2 km from the base station.
Each station within the radio communication network consists of one radio. Because of line-of-site restrictions, several stations double as repeaters throughout the current network. As of this writing, 10 of the 28 WTM stations in the radio communication network serve as repeaters. Two nonmesonet station repeaters are located in strategic areas of the radio communication network.
b. Cellular phone
In the midst of constructing a suitable radio communication network, it became evident that communication via radio would not be feasible in the most distant counties of the WTM domain. Cellular phone technology provided a short-term, alternative solution for collecting data from remote WTM stations beyond the radio communication network. Acceptable bandwidth, short connection times, and a favorable agreement with a local cellular phone company provide access to remote stations at an affordable cost. As of this writing, eight WTM stations employ cellular phone connections.
c. Landline phone
Mesonet stations using landline phone technology were not considered given the potential for recurring costs and station siting limitations. However, a cooperative effort with the local NWS Office and Southern Region Headquarters of the NWS has provided three landline phone connections as of this writing. One more station with a landline phone connection will be constructed later this year as part of the cooperative effort mentioned above. While 5-min data are available for these WTM stations, the cost of making a connection to each landline site every 5 min is prohibitive. Therefore, unless there is an active weather situation, the landline stations are only called once or twice an hour.
d. Microserial server—Internet
Other modes of communication are occasionally considered to improve the WTM communication network with regards to data transfer and to reduce recurring costs. One such improvement was to take advantage of ubiquitous Internet connections in/near the WTM domain. In a cooperative effort, a municipality in the Texas Panhandle provided an Internet connection free of charge along with tower space on the roof of their building for a radio antenna. A Campbell Scientific NL100 microserial server serves as the mesonet station’s interface to the Internet and a spread-spectrum radio connection provides a link between the microserial server and the mesonet station. Polling software on the base station computer then initiates a call to the mesonet station through the Internet and radio link. Five-min data are available on a real-time basis.
The advantages of Internet-based data collection system are numerous and might have far reaching effects that could determine the future of the WTM. Some advantages are reduction or elimination of recurring costs, no FCC licensing required for one watt spread-spectrum radio technology, and using an existing network backbone instead of having to build and/or maintain an associated network backbone.
6. Data distribution
Another goal of the WTM is to make mesonet data freely and conveniently available to a wide variety of users via the Internet. However, a computer program had to be created that would ingest, parse, and produce output in a useable format. Within a few days, the first products were made available online to the public on the WTM Web site (see below). Archived data are available upon request.
a. World Wide Web (WWW)
The first mesonet products appeared on the WTM Web site (http://www.mesonet.ttu.edu) and included basic climatological information such as high and low temperatures for the day, total liquid precipitation, peak wind speed/direction, etc. Coded meteorological observations were also made available in METAR and Standard Hydrometeorological Exchange Format (SHEF) for regional NWS Offices and the West Gulf River Forecast Center in Fort Worth, Texas. As of this writing, accessing WTM data through the Web site continues to be the most popular method for viewing and ingesting data.
b. File Transfer Protocol (FTP)
File Transfer Protocol was made available to those users who preferred this method of ingesting data as opposed to the WWW. Several user accounts were set up to meet their needs. These users preferred FTP to WWW access because of automation issues such as script writing.
c. Local Data Manager (LDM)
Unidata’s Local Data Manager is also used to distribute WTM data to several universities as well as other government agencies.
7. Quality assurance
The current configuration of the WTM produces 11 520 daily meteorological observations (288 meteorological observation sets each day from each of the 40 stations). The large number of observations requires a comprehensive quality assurance/quality control (QA/QC) procedure to check the quality of the data and provide necessary advisory information. QA/QC for the WTM is performed in two steps. In the first step, predefined tests are applied to the collected data in an effort to highlight or flag suspicious or potentially bad data. In the second step, the flagged data are examined by a “decision maker” to resolve the legitimacy of the data. Beyond these established QA/QC routines, the WTM staff provides qualitative checks on all incoming data to identify failing instruments, communication problems, or other issues affecting data quality and availability. Once identified, repairs are made as quickly as possible.
a. Data
There are two observational groups, agricultural and meteorological. The meteorological parameters are observed every 5 min and consist of 15 parameters as shown in Table 3. Agricultural parameters are observed every 15 min and consist of the 10 parameters shown in Table 4.
b. Procedures
The quality of the data is not only a concern for researchers, but also real-time users. Since the machines perform objectively by nature, QA/QC procedures become significantly important especially for datasets obtained from automated sites. Different types of QA/QC tests are proposed and applied in the literature depending on the kind of network, type of observation, and observed parameters (e.g., Wade 1987; Meek and Hatfield 1994; Gandin 1988). The QA/QC tests defined by Shafer et al. (2000) are applied to WTM datasets in much the same manner as they are applied to the Oklahoma Mesonet (Fiebrich and Crawford 2001). A FORTRAN application was developed to apply the tests; additional tests and modifications have been made for the West Texas territory, which the network serves. Critical data values required for each test were examined to ensure they were set appropriately for West Texas.
Before applying the QA/QC tests, the raw data for each site are separated into monthly agricultural and meteorological data files, providing for a convenient file size and naming convention. A standard format is incorporated that includes site ID, the time of the observations, and the parameter values. The gaps in the agricultural and meteorological observations are filled with “−99.99” values to provide consistency in the QA/QC test applications and for the benefit of future users. Given the rigorous maintenance schedule, which includes downloading all data from each datalogger, no data are lost strictly because of communication problems. Approximately 99.99% of the potential agricultural and meteorological parameters are collected.
Once collected, five distinct tests are applied to determine the quality of each single observation. For each test, the observation is flagged with a “0,” “1,” “2,” or “3” flag numbers in order to define the confidence level of the observation where “0” indicates “good,” 1 indicates “suspect,” 2 indicates “warning,” and 3 indicates “failure.” Each observation is initially assigned with the flag number 0 for each test and flag number for the corresponding test is updated, if the observation fails.
Each observation is flagged based on the result of the range test, step test, persistence test, like instrument test and spatial test. The upper/lower limitations and threshold values for each test are predefined by taking the meteorological records and instrumentation limitations (Sonmez and Doggett 2003). The range test is performed to check if the observation lies between reasonable parameter levels. Any observation outside the reasonable range is flagged as a failure. Unexpected jumps in the data are qualified with the step test. Adjacent data points with a difference higher than the predefined maximum step allowance are flagged as a warning.
The persistence test is used to examine a group of observations within a specific time window. It is useful in locating damaged instruments or those that might be “stuck” at a particular reading. The first step of the test, the variance test step, compares the variance of the dataset contained within the time window with the predefined threshold variance. All the observations in the time window are flagged as warning if the estimated variance is less than the threshold variance. The data are flagged as suspect if they are less than 2 times the threshold variance. The second step evaluates the maximum difference (difference between maximum and minimum value of the time window) with the predefined delta threshold. The observations are flagged as either warning or suspect depending on whether the maximum difference is less than the minimum delta threshold or less than one and one-half times the delta threshold. The zero test is then applied as the last step and captures continuously repeating observational values of zero. All of the observations in the time window are flagged as warning if the number of recorded zero values exceed 2.
The like instrument test is used to compare observations of like parameters. The test is performed on the temperature observations taken at 2 and 9 m and wind speed observations at 2 and 10 m. Temperature observations are flagged as suspect if the difference between two parameters is greater than a predefined threshold value. On the other hand, observations are flagged as warning if the difference is greater than 2 times the threshold value.
A one-pass Barnes analysis (Barnes 1964) is performed for the spatial test to determine the gross errors in individual observations. For this test, the observations from neighboring sites are used to get an estimated observation for a particular site. A standardized value is obtained by using the estimated and observed values for the site. The observation is flagged as warning if the standardized value is greater than 3, or suspect if the standardized value is greater than 2.
All the flag information is accumulated in an output file for the decision maker and prospective users of the dataset. At this point, the flag information and raw and corrected data files are kept completely separate, but a database combining the information is planned for the future. The flagged percentages of all meteorological and agricultural parameters based on the 2002 dataset (representative of 34 stations) are presented in Tables 5 and 6.
8. Case study/example
The WTM observed a high wind event on 15 April 2003. This significant wind event resulted in an outbreak of severe weather, including 11 tornado reports in West Texas and southwest Oklahoma, and a blinding dust storm in southeast New Mexico and West Texas. The storm was a product of a dynamic upper-level low ejecting rapidly from the desert southwest into the Southern High Plains. The progression of the upper-level low resulted in rapid surface cyclogenesis in southeast Colorado with a well-defined dryline setup south through West Texas. The dryline acted as a trigger for the initiation of a number of discrete supercells and subsequently aided in the formation of a squall line in the moist air east of its position. Following the dryline passage, winds quickly shifted to the southwest, ushering in much drier air and dust. Figure 6 displays the surface wind speed and direction throughout the WTM domain at 1730 LST. The figure shows strong south winds present in the southeast counties with winds rapidly shifting to the southwest behind the dryline in the central and western counties.
The WTM observations allowed forecasters access to timely information, thus enhancing the understanding of a rapidly evolving severe weather event. Additionally, the mesonet observations provided a good source of near real-time warning verification for the local NWS Office. On this day the WTM observed 16 severe wind reports ≥ 50 kt (25.7 m s−1) over the South Plains.
In addition to providing greatly improved resolution (both temporal and spatial) in the surface observations, the WTM also possesses a sounding system that was utilized during this event. Two soundings, launched at 1500 and 1800 UTC from Reese Center, were made readily available on the Internet. Additionally, the soundings were passed directly to the NWS. Figure 7 presents the 1800 UTC sounding. The sounding was released approximately four hours before the dryline passed through Reese Center and indicates strong south-southwest winds at the surface veering to >100 kt (51.4 m s−1) west-southwest winds aloft. The sounding also depicts a weakly capped atmosphere with marginal CAPE (394 J kg−1). Placing this vertical wind profile over the more moist south-southeast wind present in the eastern counties suggests the increased likelihood for the formation of severe convection, including the development of supercells capable of producing tornadoes.
Therefore, in this case, the presence of the WTM aided in more accurately identifying the existence, location, and evolution of the dryline as it translated through West Texas. It also provided a timely, accurate, and efficient platform for NWS to verify severe thunderstorm and high wind warnings.
9. Summary
The WTM now maintains 40 surface stations located over western Texas. Each station observes and relays (via radio, cell or land phone, or through an Internet connection) real-time (5-min) meteorological data back to Reese Technology Center in Lubbock, Texas, the base for the project. In addition to meteorological data, 15-min agricultural data are also collected and relayed. All real-time data are available and archived data can be requested. Data are distributed through a variety of methods includes WWW, FTP, and LDM. Quality control and assurance procedures have been developed. Questionable and/or bad data are automatically flagged through a set of FORTRAN routines, while a human decision maker renders a final conclusion. A Web interface to enable access to the archived dataset is currently under development.
Acknowledgments
The WTM was funded by a grant from the Texas Department of Economic Development (TDED). Additional support has been provided by the Wind Science and Engineering Research Center and Geosciences Department at Texas Tech University, and NIST (Department of Commerce NIST/TTU Cooperative Agreement Award 70NANB8H0059). This project would not be possible without the support of the local county judges, city officials, and private citizens who graciously allow us the use of their land to install mesonet stations free of charge. The authors thank Jeff Walsh and Mark Conder for developing some of the schematics and diagrams used within this article. The authors thank three anonymous reviewers who provided valuable input to improve this paper.
REFERENCES
Barnes, S. L., 1964: A technique for maximizing details in numerical weather map analysis. J. Appl. Meteor., 3 , 396–409.
Brock, F. V., Crawford K. C. , Elliott R. L. , Cuperus G. W. , Stadler S. J. , Johnson H. L. , and Eilts M. D. , 1995: The Oklahoma mesonet: A technical overview. J. Atmos. Oceanic Technol., 12 , 5–19.
Brotzge, J. A., and Crawford K. C. , 2000: Estimating sensible heat flux from the Oklahoma Mesonet. J. Appl. Meteor., 39 , 102–116.
Fiebrich, C. A., and Crawford K. C. , 2001: The impact of unique meteorological phenomena detected by the Oklahoma Mesonet and the ARS Micronet on automated quality control. Bull. Amer. Meteor. Soc., 82 , 2173–2187.
Gandin, L. S., 1988: Complex quality control of meteorological observations. Mon. Wea. Rev., 116 , 1137–1156.
Gillespie, T. J., and Kidd G. E. , 1978: Sensing duration of leaf moisture retention using electrical impedance grids. Can. J. Plant Sci., 58 , 179–187.
Meek, D. W., and Hatfield J. L. , 1994: Data quality checking for single station meteorological databases. Agric. For. Meteor., 69 , 85–109.
Richardson, S. J., 1995: Multiplate radiation shields: Investigating radiational heating errors. Ph.D. dissertation, University of Oklahoma, 134 pp. [Available from 100 E. Boyd, Suite 1210, Norman, OK 73019.].
Shafer, M. A., Hughes T. , and Carlson J. D. , 1993: The Oklahoma mesonet: Site selection and layout. Preprints,. Eighth Symp. on Meteorological Observations and Instrumentation, Anaheim, CA, Amer. Meteor. Soc., 231–236.
Shafer, M. A., Fiebrich C. A. , Arndt D. S. , Fredrickson S. E. , and Hughes T. W. , 2000: Quality assurance procedures in the Oklahoma Mesonetwork. J. Atmos. Oceanic Technol., 17 , 474–494.
Sonmez, I., and Doggett A. L. , 2003: The West Texas data archive: A tool for wind science and engineering research in West Texas. Proc. 11th Int. Conf. on Wind Engineering, Lubbock, TX, ICWE/NIST, 2021–2028.
Wade, C. G., 1987: A quality control program for surface mesometeorological data. J. Atmos. Oceanic Technol., 4 , 435–453.
World Meteorological Organization, 1983: Guide to Meteorological Instruments and Methods of Observation. 5th ed. World Meteorological Organization, 483 pp.
Map of West Texas Mesonet stations with WTM four-letter IDs.
Citation: Journal of Atmospheric and Oceanic Technology 22, 2; 10.1175/JTECH-1690.1
Elevation view of a typical West Texas Mesonet station. The 2-m temperature probe is not easily depicted given its location and has been omitted. Please see the plan view (Fig. 3) for more information.
Citation: Journal of Atmospheric and Oceanic Technology 22, 2; 10.1175/JTECH-1690.1
Plan view of a typical West Texas Mesonet station.
Citation: Journal of Atmospheric and Oceanic Technology 22, 2; 10.1175/JTECH-1690.1
Photograph of the West Texas Mesonet site at Snyder 3E (SNYD).
Citation: Journal of Atmospheric and Oceanic Technology 22, 2; 10.1175/JTECH-1690.1
West Texas Mesonet data flow from remote stations to users.
Citation: Journal of Atmospheric and Oceanic Technology 22, 2; 10.1175/JTECH-1690.1
Wind speed and direction for the West Texas Mesonet at 1730 LST 15 Apr 2003. One full barb represents 5 m s−1 and a flag represents 25 m s−1.
Citation: Journal of Atmospheric and Oceanic Technology 22, 2; 10.1175/JTECH-1690.1
Sounding launched from Reese Center at 1800 UTC 15 Apr 2003. The temperature (solid line) and dewpoint (dashed line) are in degrees Celsius. The wind speed is in knots.
Citation: Journal of Atmospheric and Oceanic Technology 22, 2; 10.1175/JTECH-1690.1
West Texas Mesonet equipment/sensors.
West Texas Mesonet sensor characteristics, accuracy, and data resolution.
Meteorological parameters recorded using a 5-min temporal resolution.
Agricultural parameters recorded using a 15-min temporal resolution.
Percentage of flagged meteorological data by parameter based on the 2002 dataset.
Percentage of flagged agricultural data by parameter based on the 2002 dataset.
