Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
4 Observing Systems and Technologies: Successes and Challenges In this chapter we highlight current and emerging observing systems and technologies to address the observational needs discussed in Chapters 2 and 3. Systems and technologies are considered in two broad categories: those based on the surface and those based in space. In the spirit of consid- ering the issue From the Ground Up, those based at the surface are given more emphasis and further categorized according to whether the technol- ogy provides in-situ or remotely sensed observations. Surface-based remote sensing systems are discussed according to whether the sensing technology is active or passive. We discuss systems that may be based at the surface but provide both in-situ and remotely sensed observations in the vertical dimen- sion at heights well above near-surface. Some of these systems are mobile (e.g., aircraft). Others are designed to provide targeted observations. Following the discussion of technologies and systems, we summarize several particular observational challenges, including those posed by the surface and the planetary boundary layer, and mountains, cities, and coasts. We conclude the chapter with a discussion of the global context within which U.S. mesoscale observations are embedded. The global context is important because, for many applications, the utility of limited-area meso- scale observations is highly dependent on larger domains of observations, for example, in the provision of initial and boundary conditions for meso- scale numerical weather prediction models. 87
88 OBSERVING WEATHER AND CLIMATE FROM THE GROUND UP SURFACE-BASED OBSERVING SYSTEMS Mesoscale meteorology is closely identified with surface observing sys- tems, perhaps because disruptive weather is intrinsically at the mesoscale, and impacts are most often experienced at or near the surface. The United States has enormous diversity and complexity within its inventory of sur- face-based observing assets, which are operated by federal, state, and local agencies, numerous segments of the private sector, universities, schools, and hobbyists and other enthusiasts. Surface-based observing systems employ both in-situ sensing as well as active and passive remote sensing technolo- gies. A number of efforts have summarized observational capabilities in the United States. For the last decade and with funding from the Global Energy and Water Cycle Experiment (GEWEX) Americaâs Prediction ÂProject (GAPP), University Corporation for Atmospheric Research/National Center for Atmospheric Research (UCAR/NCAR) has developed a database that describes and maps what is available (http://www.eol.ucar.edu/projects/ hydrometnet). The National Science Foundation (NSF) recently sponsored development of another database to serve the dual purpose of providing users with information about available resources and to identify future observational needs in atmospheric research (see http://www.eol.ucar.edu/ fadb/). The National Oceanic and Atmospheric Administration (NOAA) is currently developing an Observing Systems Architecture website with a comprehensive list of NOAA networks at http://www.nosa.noaa.gov (check âObserving System Inventoryâ on the left side of the page). A summary table based on these websites appears in Appendix B. Other useful websites for such information include http://madis.noaa.gov and http://www.met. utah.edu/cgi-bin/databbase/mnet_no.cgi. Networks for Surface Observations: Land-Based Most commonly, âsurfaceâ measurements consist of temperature and relative humidity, wind, precipitation, and air pressure. World Meteoro- logical Organization (WMO) standards prescribe wind measurements at a height of 10 m in open areas, and pressure, temperature, and humidity at about eye level (1.5 m), but many surface measurements deviate from these standards, often for good reason. For example, routine observations are made for applications in transportation, agriculture, the power industry, air quality, and public safety, nearly all of which have specific criteria that differ from WMO standards. There are many thousands of surface sites gathering weather and related information. Based on the UCAR/NCAR and NSF surveys, approximately 500 surface networks operate in the United States and its coastal waters. Federal and state agencies as well as universities and the private sector take
OBSERVING SYSTEMS AND TECHNOLOGIES 89 observations off the coasts of the lower 48 states, Alaska, and Hawaii. The federal government alone operates approximately 25,000 sites for numer- ous applications, including climate monitoring, weather forecasting, and monitoring conditions near fires; however, many sites do not report data in real time. Many state departments of transportation, working with private- sector transportation weather service providers, operate networks along highways, and at least one railroad collects observations along its tracks. States, cities, and universities maintain mesoscale networks for air quality monitoring, as part of their flash-flood warning procedures, for agriculture, research, and general weather information. A relatively recent develop- ment is urban networks for use in the case of deliberate or accidental toxic releases. Other groups that collect data are power companies, chemical pro- cessing plants, and television stations. Even private citizens have automated weather stations in their homes, some of which produce real-time data. Although surface sites are numerous, abundance doesnât necessarily translate into utility. The sites are not evenly distributed: There are gaps in rural areas, areas with limited access, and in complex terrain. It is a major effort to keep information on multiple networks up to date, so some of net- works included in Appendix Table B.1 may have languished due to lack of funds: The numbers are always changing. On the other hand, some smaller networks may not be documented. Figure 4.1 maps the surface coverage for meteorological data over Washington State. The data are from NorthwestNet, which collects and integrates measurements made by multiple groups. On the map, one can see areas of dense and sparse data coverage. The latter areas typically have low population density or are difficult to reach due to terrain or other factors. The densely covered areas have data from multiple sources, including weather hobbyists, air pollution networks, and road networks, as well as more conventional sources. Not all of obser- vations are suitable for all applications. For example, roadside weather stations are installed along stretches of road with frequent hazardous weather, such as high winds or icing, so they are often not ârepresenta- tiveâ of synoptic conditions. However, such non-representativeness on the Âsynoptic scale is strong evidence of value at the mesoscale and is the primary driver for such extensive private and public investment in surface stations nationwide. Likewise, data from individual homes and schools may not meet the accuracy standards or exposure criteria that are required for numerical weather prediction or research. However, nearly all observations are suit- able for some purposes, such as identifying the passage of a strong front with a well-defined wind and temperature change. It is possible that some specialized networks have sites with higher quality data than one would expect, but the supporting metadata are absent.
90 OBSERVING WEATHER AND CLIMATE FROM THE GROUND UP FIGURE 4.1â Sample map of NorthwestNet surface observations. SOURCE: Figure 4-1.eps provided courtesy of Cliff Mass, University of Washington. bitmap image While the technology for many of the âweatherâ variables is mature, measurement of rainfall and especially surface snowfall and precipitation type remains a challenge. Rainfall measurements from gauges are reason- ably accurate, but rainfall varies on scales smaller than the typical spacing between gauges; this problem has been alleviated to some degree by com- bining gauge and weather radar measurements. The Natural Resources Conservation Center operates the Snowpack Telemetry (Snotel) network of âsnow pillowsâ that weigh the snow using pressure sensors to estimate the water supply. Data are routinely available daily but are accessed at higher rates for special needs. The type and amount of frozen precipitation is criti-
OBSERVING SYSTEMS AND TECHNOLOGIES 91 cal to keeping the roadways passable (NRC, 2004a). Precipitation type and snowfall rate are critical information at airports. While networks that collect weather data can be dense, as illustrated in Figure 4.1, networks that collect soil moisture can seem sparse by com- parison, as illustrated in Figure 4.2. A notable exception is the Oklahoma Mesonet (see Box 4.1). Soil-moisture estimates are relevant for numerical prediction and agricultural applications, among others. Automated tech- niques exploit, for example, the variation of the dielectric constant for soils (time-domain reflectometry), neutron scatter by water in the soil (neutron probes), and measuring how a ceramic block embedded in the soil reacts to heat pulse. The dearth of soil-moisture data is currently being addressed by running land-surface models that integrate precipitation, solar radiation, etc., for a period of time. Satellites, to be discussed in the next section, have potential to supply near-surface soil-moisture data, but estimates are limited by clouds and thick vegetation. Larson et al. (2008) have suggested a new technique for tracking soil-moisture fluctuations that is independent of FIGURE 4.2â Soil-moisture networks in the United States documented at http:// 4-2.eps www.eol.ucar.edu/fadb/. NOTES: The black dots represent the Oklahoma Mesonet; bitmap image green, the Illinois State Water Survey network; yellow, ARM/CART; white, Ameri- Flux sites; red, United States Department of Agriculture/Natural Resources Conser- vation Service (USDA/NRCS) Soil Climate Analysis Network (SCAN). SOURCE: Courtesy of Scot Loehrer.
92 OBSERVING WEATHER AND CLIMATE FROM THE GROUND UP BOX 4.1 The Oklahoma Mesonet The most prominent state mesonet is the Oklahoma mesonet (Figure 4.1.1), which is used for emergency response, agriculture, severe storms forecasting, reÂ search, and other applications (McPherson et al., 2007). The Oklahoma mesonet consists of 120 automated stations, with at least 1 station in each of Oklahomaâs 77 counties. At each site, the environmental variables are measured by a set of instruments located on or near a 10-m-tall tower. The Oklahoma Climatological Survey (OCS) at the University of Oklahoma receives the observations, verifies the quality of the data, and provides the data to mesonet customers. It takes only 5 minutes to make measurements available to the public. FIGURE 4.1.1â Map of the Oklahoma mesonet. 4-3.eps NOTE: Multiple agencies are involved in the individual sites. bitmap image The standard measurements include temperature and humidity (1.5 m), wind (10 m), air pressure, precipitation, incoming solar radiation, and soil temperature at 10 cm either below natural cover or bare ground. Most sites also sample air temperature at 9 m above ground, wind speed at 2 and 9 m above ground, soil moisture at 5, 25, and, 60 cm below ground, soil temperatures at 5 and 30 cm below ground under the natural sod cover, and soil temperature at 5 cm below bare ground. At 10 sites, turbulence fluxes of heat, moisture, and momentum are sampled at half-hour intervals in addition to the soil and weather variables.
OBSERVING SYSTEMS AND TECHNOLOGIES 93 cloudiness; it exploits the effect of soil moisture on the reflection of Global Positioning System (GPS) radio waves. Currently, available remote sensing technologies cannot provide soil moisture below approximately 5 cm. In addition to numerical data, there is a growing network of web cameras monitoring the nationâs streets and highways. While not especially useful for numerical weather prediction, cameras are highly useful for road trans- portation, providing drivers and road managers a check on road conditions (weather, traffic flow, state of the road due to precipitation), and for moni- toring wind and weather changes in other applications, such as fighting forest fires or warning about the spread of noxious substances. Coastal Ocean Networks The Integrated Ocean Observing System (IOOS) provides real-time quality-controlled data for both the oceans and the Great Lakes, âfrom the global scale of ocean basins to local scales of coastal ecosystems.â IOOS is an end-to-end system that involves observations, data communications and management, and data-analysis and modeling, through its three inter- acting subsystems, Observation and Data Telemetry, Data Management and Communications, and Data Analysis and Modeling. These challenging tasks involve partnerships among federal and state agencies, the private sector, and universities. IOOS has a coastal component, which involves the U.S. Exclusive Economic Zone (EEZ, which extends 200 nautical miles or 370 km offshore) and the Great Lakes, and a global component. Coastal and interior waters in the United States are monitored by a diverse network of buoys operated by both the public and private sectors. These diverse measurements are being incorporated into 11 Regional Coastal Ocean Observing Systems (RCOOSs), parts of which also participate in a National Backbone of coastal observations. Most of the RCOOS buoys measure meteorological variables. NOAAâs National Data Buoy Center col- lects and quality-checks, and then distributes the data via the GTS in real time. The core variables measured by the National Backbone sites include ocean data on composition (salinity, dissolved nutrients, dissolved oxygen, chemical contaminants), life (fish species and abundance, zooplankton and phytoplankton species and abundance, waterborne pathogens), and other physical characteristics (temperature, sea level, surface waves and currents, heat flux, bathymetry and bottom character, sea ice, optical properties). The RCOOSs (Figure 4.3) are being coordinated by regional associa- tions that will in turn contribute to the evolving IOOS. There are of the order of 700 coastal observation sites in approxi- mately 50 networks. Since these sites must cover the Great Lakes and the â See http://www.ocean.us/what_is_ios.
94 OBSERVING WEATHER AND CLIMATE FROM THE GROUND UP 4-4.eps FIGURE 4.3â Regional Coastal Observing Systems. NOTE: LME = Large Marine Ecosystems. SOURCE: Nationalbitmap image http://www.ndbc.noaa.gov/. Data Buoy Center, U.S. coastline plus the EEZ, the coverage is sparse compared to the land surface. To illustrate, the continental United States has a surface area of 7,700,000 km2, while a conservative estimate of the EEZ is slightly less than one-third that value. As shown in Appendix Table B.1, the number of meteorological sites on land reporting in real time exceeds 10,000. Signifi- cant deficiencies exist over the coastal waters despite the fact that oceanic regions tend toward greater uniformity over larger regions. The 700 coastal ocean sites, which include Alaska and Hawaii, clearly do not resolve either the atmospheric or oceanic mesoscale (Figures 4.4 and 4.5). To counter this large difference in the density of surface observations between land and sea, satellite scatterometer winds and sea-surface tem- perature estimates provide high-quality information at high resolution over the oceans. However, some of these measurements become problematic very close to the coasts, owing to strong gradients and land-contaminated satel- lite footprints. The low density of measurements immediately offshore is a matter of considerable concern, given that 50 percent of the U.S. popula- tion lives within 50 miles of the coast and the increased complexity and importance associated with coastal airflow near large cities.
OBSERVING SYSTEMS AND TECHNOLOGIES 95 FIGURE 4.4â Coastal networks along U.S. Pacific coastlines. 4-5.eps bitmap image 4-6.eps FIGURE 4.5â Sites in the Pacific Northwest. SOURCES: GAPP/NCAR Earth O bitmap image Â bserving Laboratory, http://www.eol.ucar.edu/projects/hydrometnet. Figure from National Data Buoy Center, http://www.ndbc.noaa.gov/.
96 OBSERVING WEATHER AND CLIMATE FROM THE GROUND UP The Vertical Dimension: Surface-Based In-Situ Technologies The high cost of atmospheric measurements relative to surface mea- surements is the reason why the government bears the cost of many obser- vations taken well above ground. The one major system, the radiosonde network, launched from the ground at fixed times to collect observations at various altitudes, is described below. Radiosondes are balloon-borne instrument packages that measure temperature, relative humidity, and wind as a function of pressure, from near the surface to stratospheric altitudes (generally 10 hPa or higher). Radiosondes have been the standard for atmospheric measurements in the troposphere and lower stratosphere since World War II. The vertical resolu- tion of the measurements is good, better than 10 hPa. The roughly 80 U.S. raob sites are widely spaced, several hundred kilometers apart. Balloons are u Â sually launched twice a day at 0000 and 1200 UTC. Thus, the sampling density of the radiosonde network is poorly matched to the large amplitudes and small scales of lower tropospheric variability. In 2000, the National Research Council (NRC) Panel on Geoscience, Environ- ment, and Resources discussed improved temperature monitoring capabili- ties from this network (NRC, 2000). It found the radiosonde network to be in decline and insufficient even for global monitoring. The trend has persisted and is not likely to reverse. In addition to the standard temperature, relative humidity, and wind radiosonde data, a subset of the U.S. network contributes to the WMOâs Global Atmosphere Watch Ozone monitoring network. With ozonosondes launched in tandem with a modified radiosonde, telemetered ozone profiles are available at approximately 100 sites globally. These data are important for stratospheric ozone, but the slow sensor response limits the applica- tion of such profiles in lower tropospheric applications unless balloons with a slower rise rate are used or inexpensive fast-response sensors are developed. Evolution of the technology associated with disposable sondes conÂ tinues; the dollar cost per sounding has declined, and the quality of data continues to improve. Small disposable nanosensors are currently being tested, which may make more parameters (trace gases, for example) pos- sible to measure from profiling sondes. Further development of sensor technology for carbon dioxide, ozone, and other priority pollutants is encouraged since the technology to profile these variables via remote sens- ing is not mature or sufficiently cost-effective. This need for obtaining pro- files for âchemical weatherâ variables described in the âDecadal Surveyâ (NRC, 2007a) would suggest that the current network be maintained as a source for profiling information for additional variables.
OBSERVING SYSTEMS AND TECHNOLOGIES 97 The Vertical Dimension: Surface-Based Remote Sensing Technologies Passive and active remotesensing techniques have been employed in up- looking configurations at the surface. As examples, several types of sensors are described below that utilize microwave, infrared, and visible parts of the electromagnetic spectrum. Passive Sensors Microwave radiometers. The microwave spectrum from a few to 180 GHz frequency contains a wealth of information on water and hydrometeors in the atmosphere. Outside of a broad O2 absorption feature at 60 MHz, the spectrum is dominated by the pressure- and temperature-dependent spec- trum of water vapor, liquid water, and ice. Up-looking microwave spectro- radiometry has been used to retrieve profiles of temperature, water vapor, and cloud liquid water (Solheim et al., 1998). The temporal resolution of the profiles is excellentâ5 minutesâbut the vertical resolution decreases quickly with altitude and is coarser than that for radiosondes. Several Atmospheric Radiation Measurement/Clouds and Radia- tion Testbed (ARM/CART) sites operate microwave radiometer profilers (MWRP) that measure downwelling microwave radiation in two frequency ranges: 22-30 GHz and 51-59 GHz (Liljegren, 2007). The former range contains a weakly absorbing water vapor resonance band; measurements in five channels are used to infer water vapor profiles. The latter frequency range lies on one shoulder of the broad oxygen absorption band mentioned above. Measurements in seven channels are used to infer temperature pro- files. The profiles, along with cloud liquid water path are derived at roughly 5-minute intervals. GPS Integrated Precipitable Water. An analysis of GPS signal delays that result from the radio refractive index profile leads to estimates of Â(columnar) Integrated Precipitable Water (IPW; Bevis et al., 1992). IPW indicates the depth of liquid water that would result if all water vapor in a vertical col- umn were condensed. Except during maneuvers of GPS satellites, IPW esti- mates are stable, accurate except during intense rainfall, and do not need calibration. GPS/IPW measurements are thus used as a reference standard to calibrate rawinsondes There are 300 to 400 ground-based receivers in the United States that report hourly (Figure 4.6). Most of the GPS receiver sites in the United States were in place before it was recognized that water vapor, a nuisance for geodetic applications, was producing a useful sig- nal for atmospheric applications. Both IPW and slant-path water vapor â For more information, see http://www.arm.gov/instruments/instrument.php?id=mwrp.
98 FIGURE 4.6â The surface-based GPS network in the United States. 4-7.eps 2 bitmap images BROADSIDE
OBSERVING SYSTEMS AND TECHNOLOGIES 99 measurements (Braun et al., 2003) have been found useful for analyzing atmospheric water vapor content. SuomiNet (Ware et al., 2000), an array of surface-based GPS receivers, includes a mesoscale (spacing 50-60 km) array in Oklahoma, where it can be ingested into experimental Numerical Weather Prediction (NWP) models used to predict severe storms. Advanced Emitted Radiance Interferometer. In the infrared, many more gases (CO2, CO, O3, CH4, NO, H2O, etc) have spectral features that can be used for ground-based atmospheric profiling. The Atmospheric ÂEmitted Radiance Interferometer (AERI) was developed at the University of ÂWisconsin and now has been more widely distributed in variants using commercial Fourier trans- form interferometers. The AERI has approximately 1 cmâ1 spectral resolution across much of the infrared spectrum. Main products of the instrument are temperature and water vapor profiles that use an infrared retrieval of multiple wavelengths near 15 Î¼m, 4.3 Î¼m, and 6.6 Î¼m. Additionally, CO columns have been measured and have been used as indicators of biomass burning (He et al., 2001). Results for O3, NO2, CH4, and fluorocarbons by AERI radiances have been reported by Evans et al. (2002). Sun photometry. A number of networks exist globally for the determination of aerosol optical depth using the attenuation of direct sunlight. At least three technologies are able to make these measurements at this time: (1) the Physikalisch-Meteorologisches Observatorium Davos (PMOD) instrument used in a series of remote observation sites in the Global Atmosphere Watch (GAW) aerosols network (FrÃ¶hlich et al., 1995); (2) the shadow- band Âradiometer in a number of confederated networks (the United States Department of Agriculture UV-B network, the Michalsky Network, the NOAA Baseline Surface Radiation network, and the Surface Radiation network); and (3) the Cimel sunphotometer (CSPHOT) as part of the NASA Aerosol Robotic Network (AERONET; Holben et al., 2001) and the ÂPHOTONS network. A recent review of networks of aerosol optical depths has been presented by the WMO GAW (2004). Sun photometer net- works used in conjunction with satellite optical depth measurements have been used for spatial and temporal extinction of ground-based particulate mass measurements (Engel-Cox et al., 2006). Active Sensors Scanning radars. When active remote sensors such as radars are able to perform scan sequences on time scales commensurate with mesoscale â PHysics, Optoelectronics, and Technology Of Novel Microresonator Structures.
100 OBSERVING WEATHER AND CLIMATE FROM THE GROUND UP atmospheric evolution, volumetric snapshots of atmospheric structure may be obtained. Volumetric structure is often key to identification of severe storms, rain changing to snow, hail versus rain, pollution plume rise versus fumigation, etc. Weather radars are sensitive to precipitation and also insects, birds, and refractive-index gradients in clear air. The U.S. WSR-88D network of over 150 10-cm scanning Doppler radars (Figure 4.7) is essential for detecting and tracking storms of all kinds, including severe storms, and the related issuance of public warnings. The Federal Aviation Administration (FAA) operates 47 Terminal Doppler Weather Radars near major U.S. airports to detect and report hazardous weather around airports. In precipitation, the radial velocity data are useful to estimate the mesoscale wind field, espe- cially embedded rotation or convergence. Clear-air echoes can be used to estimate the boundary-layer wind field through Velocity-Azimuth Display (VAD) techniques or echo tracking. Because backscatter from insects is significant in otherwise clear air, such information is typically available for the boundary layer when the temperature is above 10 Â°C. Soon, WSR-88D radars will be equipped with polarimetric capability, which will improve estimates of precipitation amount and type. Television stations in many markets operate Doppler radars in competition for viewership as part of their weather broadcasts; data are sometimes shared with local weather service offices for severe-storm nowcasting. Other radars (e.g., the CHILL radar operated by Colorado State University) and the S-Pol radar operated by the National Center for Atmospheric Research are used for research. Even though the WSR-88D network is a core source of mesoscale meteorological information, its ability to give high-spatial and -temporal information has limitations, as discussed in an NRC report (NRC, 2002). One of the limitations is related to the spreading of the beam with increas- ing distance from the radar and the curvature of the Earthâs surface. At 0Â° elevation angle, because of the Earthâs curvature, the volume being probed at long ranges can be at several kilometers height and averaged over several kilometers depth (Figure 4.7). In winter, the 0Â° elevation beam overshoots shallow precipitating clouds that are producing snowfall, and, when radars are located on hilltops near the West Coast, clouds producing substantial âwarm-rainâ precipitation. For systems designed to probe boundary-layer phenomena, this limits the range of applicability of radar to 100 km or less. Given these simple geometric facts it comes as no surprise that WSR-88D coverage is discontinuous, having been designed to satisfy a 3 km (10,000 ft) constant altitude specification. One possible solution to this deficiency is to deploy additional radars in much greater numbers. If such radars were of the WSR-88D (10 cm) type, the cost would be quite high, and the radar coverage would be excessively duplicative in other applications. The Collaborative Adaptive Sensing of
OBSERVING SYSTEMS AND TECHNOLOGIES 101 FIGURE 4.7â Current U.S. WSR-88D coverage and the gap in coverage generated by 4-8.eps the Earthâs curvature. SOURCE: McLaughlin (2005), figure from the Collaborative Adaptive Sensing of the Atmosphere 2007images the committee. 2 bitmap briefing to
102 OBSERVING WEATHER AND CLIMATE FROM THE GROUND UP the Atmosphere (CASA) project has been developed at the initiative of the National Science Foundation. CASA is currently evaluating the effective- ness of small (3 cm) prototype radars in testbed mode, as if these were to be densely distributed across the United States (McLaughlin et al., 2005). Potentially thousands of such radars could be placed on buildings and cell phone towers. CASA has adaptive scanning technologies that are designed to intelligently seek targets of interest (mesovortices, for example) and to increase the sampling of these targets close to the surface. A 2007 experiment with these types of systems was conducted in Okla- homa and provided a striking example of the need for this technology. On May 9, 2007, a shallow mesovortex developed into an F1 tornado near Lawton, Oklahoma. The image (Figure 4.8) shows the hook echo prior to forming into a tornado, at very low altitude and high resolutionâobservations made pos- sible with the CASA network. The same storm produced an F1 tornado that touched down near Minco, Oklahoma, at 0350 UTC that was not detected by the WSR-88D system but was detected by forecasters at the National Weather Service using targeted CASA observations. Damage sur- veys subsequently confirmed the tornado event. Caution must be exercised in consideration of such solutions since high-frequency radar signals attenu- ate rapidly in precipitation, and backscatter easily becomes non-Rayleigh, which introduces other complications. In principle, however, multiple view angles and polarimetric methods can mitigate these complications. Other radar developments look at changes in radio refractivity of the atmosphere driven mainly by relative humidity (Fabry, 2004). This tech- nology shows promise based on tests at National Center for Atmospheric Research and elsewhere (Weckwerth et al, 2005 (JAM 44(3)). These radar systems may aid in the mapping of moisture fields in a dense, highly distrib- uted radar network. In combination with other passive technologies above in a testbed, vertical profile information on temperature and Â humidity should be widely available across the country. Cloud radars. Cloud radars operate at shorter wavelengths, from Âmillimeters to a centimeter or two. These radars paint a three-dimensional picture of multiple cloud layers, but attenuation by precipitation limits their useful range. Cloud radars are used to research clouds, tornadoes, and the clear- air boundary layer. Radar wind profilers. The introduction of radar wind profilers, operating variously at 50, 404, 449, and 915 MHz frequency, has been an important development from NOAAâs Earth Systems Research Laboratory. Horizontal and vertical winds are estimated from backscatter associated with radio refractive index gradients when measured at different beam pointing angles.
OBSERVING SYSTEMS AND TECHNOLOGIES 103 4-9.eps FIGURE 4.8â The Lawton tornado showing the hook echo (red circle, right panel) at an altitude of approximately 400 m above ground level. NOTE: The image on bitmap image the right has been filtered to remove ground clutter. ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ SOURCE: V. Chandrasakar, CSU/CASA. The sharp drop in signal at the top of the convective boundary layer is used to estimate its depth; however, estimates of the depth of the night-time boundary layer are more difficult due to its shallowness. Boundary-layer depth is a major source of uncertainty in the predictive capability of current numerical chemical forecast models. The National Profiler Network (NPN), with 32 sites in the Âcentral United States and 3 sites in Alaska, delivers wind profiles up to 17 km. Sample data from an NPN site at Conway, Missouri, are shown in Figure 4.9. Wind speed and direction are retrievable throughout the Â troposphere at good temporal and vertical resolution. At this time, data from approximately 100 Cooperative Agency Profiler (CAP) sites from over 35 different agencies from around the world are being acquired by Earth System Research Laboratoryâs (ESRLâs) Global Systems Division. The CAP sites are home to boundary-layer profilers (BLPs), small, relatively low-cost ultra high frequency (UHF) Doppler radars used primar- ily to measure vertical profiles of horizontal winds. BLPs have a minimum range of approximately 100 m above ground level (AGL) and range reso- lutions selectable from 60-400 m. Depending on the configuration of the radar and the atmospheric conditions, BLPs are capable of measuring wind up to approximately 1-5 km AGL. â See http://www.profiler.noaa.gov/npn/profiler.jsp. 4-9.eps bitmap image enlarged for broadside
104 OBSERVING WEATHER AND CLIMATE FROM THE GROUND UP FIGURE 4.9â Eleven hours of wind data from the NPN profiler at Conway, ÂMissouri. Note: The short wind barbs, long4-10.eps flags stand for 5, 10, and 50 m sâ1, barbs, and respectively. bitmap image The NPN is a nationally funded enterprise. The diverse agencies that fund the CAP are loosely affiliated, if at all; they are willing to be part of the collective that gathers the data into a single processing center. There are undoubtedly some profilers in the United States that do not belong to either network. Sodars. Along with radar wind profiling, acoustic remote sensing or sodar technology is a way to determine boundary-layer height. In a sodar, pulses of sound are emitted vertically along a single axis (and sometimes addi- tional axes at angles to the vertical) at a frequency near 1600 Hz. Sound waves in the atmosphere propagate at a known speed (a function of virtual temperature) and reflect from inhomogeneities in the density structure of the atmosphere, prevalent at temperature inversions. From the time of flight of the emitted pulse and the returned reflected sound, one can determine range to the height of the planetary boundary layer (PBL). Commercial sodars have been available since the 1970âs. The power-aperture product of the sodarâs transmitter and receiver determines the altitude to which the top of the PBL can be detected. Some systems can probe to several kilometers altitude, but they can be quite loud and annoying to nearby residents, which limits their use around populated sites.
OBSERVING SYSTEMS AND TECHNOLOGIES 105 Radio Acoustic Sounding System (RASS). Many NPN profilers are collo- cated with an acoustic transponder, hence the term Radio Acoustic Sounding System (Neiman et al., 1992). The transponder emits sound waves, whose propagation speed is detected by the Doppler radar at various ranges, thus enabling estimates of the virtual temperature profile. Eleven sites near the center of the NPN array have RASS capability to help with weather analysis and forecasting. Measurements typically extend up to 2-3 km, higher under light wind conditions. Many BLPs are also equipped with RASS. This tech- nology has the potential to estimate PBL depth under some circumstances, but, like sodar, the noise it produces can be irritating. Ceilometry. Another old technology is the use of light scattering at the base of clouds to determine ceiling heights. As early as the 1940s rotating mirror- lamp combinations were used to determine ceiling heights. Since the 1980s, small pulsed lasers have been used to determine these heights. In the light detection and ranging (lidar) configuration, a pulse of light is sent into the atmosphere. The interval of time between the emission and the detection of reflected light from the cloud determines the height of the cloud. A visible light analog to radar, lidar has been used for detecting not only cloud base but also atmospheric constituents in cloud-free air. For clouds, the system need not have high power, because the sig- nal from clouds is large. To measure ceiling height, commercial laser c Â eilometers use gallium-arsenide (Ga-As) diode lasers in the near infrared with telescopes with apertures of a few centimeters to send pulses of light into the air. For airport observations, ceiling heights of <10,000 ft are most important; thus early ceilometers focused on the measurement of lower clouds. Interest in the detection of higher clouds is growing, and the cur- rent Vaisala Âceilometers can reach to 25,000 or 35,000 ft. Highly sensitive ceilometers are also able to detect aerosols at lower altitudes. Recent work has shown that ceilometers may give information about PBL structure (and PBL height) as an additional capability. With nearly 180 ceilometers in the Automated Weather Observing System (AWOS), ceilometers may also allow derivation of aerosol profiles. To date, however, comparison of the diverse outputs of these systems has not been quality assured for aerosols or for PBL variables. Further work in this area is indicated. Lidar. A visible- or infrared- wavelength analog to radar, lidar has been used for detecting aerosols and for using aerosol backscatter to estimate PBL depth. Lidars are also used to detect trace gases. Some infrared lidars are safe to the human eye, easily operated in heavily populated areas such as Washington D.C., and capable of detecting some types of pollution and other particulate matter. The first lidars were built nearly 40 years ago, yet
106 OBSERVING WEATHER AND CLIMATE FROM THE GROUND UP the technology continues to mature rapidly and become increasingly useful. A review of lidar technology has been recently published (Weitkamp, 2005). The WMO is attempting to create the Global Aerosol Lidar Observation Network (BÃ¶senberg and Hoff, 2008). Lidar technology can be divided into elastic (i.e., systems that measure at the frequency of the emitted light) and non-elastic systems. In elastic systems, one fundamental difficulty is that the energy returned to the lidar is a function of the scattering coefficient of the scatterers (aerosols or gases) times the integrated two-way transmittance of the atmosphere out to the target and back. In radar, this attenuation is small and ignored, and the sig- nal is proportional to the scattering cross-section of the targets. In lidar, the two-way transmittance cannot be ignored, even for clear air, since ÂRayleigh scattering in visible wavelengths is significant. This forms a problem of having one measured variable (the returned energy) and two unknowns (the backscatter coefficient and the two way transmittance). To rectify this problem, Rayleigh scattering is assumed to retrieve either backscatter or extinction profiles. Precise measurements of extinction profiles are depen- dent on an assumed microphysical model in elastic lidar. Elastic lidars are simple to operate, and many networks that give pro- files of aerosols exist worldwide. In the lowest power configuration, a very high repetition rate (>2000 Hz) Nd-YLF crystal lidar oscillator is used, in conjunction with a 10-20 cm aperture receiver, to form a âmicropulseâ lidar. The network configuration of these lidars is called the Micro-pulse Lidar Network (MPL-NET), and five such systems exist in the United States. In addition, another 10-20 higher power elastic lidars exist at uni- versities and government research labs. These systems have been formed into an informal network called REALM (the Regional East Atmospheric Lidar Mesonet; Figure 4.10). Next in complication in the lidar family is the non-elastic Raman lidar technique. In this technology, the outgoing pulse at the fundamental wave- length undergoes Raman scattering from gases in the atmosphere (N2, H2O are the most common Raman-shifted frequencies monitored). The ratio of the water vapor Raman signal to the nitrogen Raman signal gives the water vapor mixing ratio as a direct result. This allows determination of relative humidity profiles, if a temperature profile is available. Similarly, the ratio of the elastic aerosol channel signal to the Raman nitrogen signal gives the aerosol mixing ratio, and, by taking a derivative of this ratio with range, the extinction of the aerosol can be measured precisely with no assump- tions about the aerosol (in contrast to the elastic systems discussed above). The Raman lidar technique has been widely used in Europe with a network called EARLINET (European Aerosol Research Lidar Network; Matthias et al., 2004) and nearly 100 journal publications have resulted from this work. In the United States, there are less than five such systems. However,
FIGURE 4.10â Simultaneous observations of lidar planetary boundary-layer heights from four REALM lidar stations from Virginia to New York City on December 4, 2006. 4-11.eps 107 bitmap image enlarged for broadside
108 OBSERVING WEATHER AND CLIMATE FROM THE GROUND UP the United States has arguably the longest continuing record from such a system at the Department of EnergyâARM/CART Raman Lidar (Turner et al., 2002) Other technologies with lidar include Doppler lidar for winds (Grund et al., 2001), differential absorption lidar (Ancellet and Ravetta, 2005) for ozone, water vapor, and nitrogen dioxide profiling; high spectral reso- lution lidar (Piironen and Eloranta, 1994); Rayleigh temperature profil- ing (McGee et al, 1995); and multi-wavelength lidar (Veselovskii et al., 2004) for aerosol microphysical characterization. Several companies (e.g., Halo, ÂLeosphere, Coherent) are producing low-power, high-repetition-rate 1.55 Âµm fiber laser transmitter Doppler lidars, which can obtain wind information to heights of several hundred meters. However, relatively few types of lidars are currently used extensively in operations because of the expense and eye-safety issues. Surface-Based Transportable and Mobile Observing Systems Not all observational networks are at a fixed location. Increasingly, commercial aircraft are instrumented. Furthermore, a number of dedicated observational and research facilities in the United States are either trans- portable (typically shipped to a location for an observational field program) or mobile (i.e., easily movable to a new location in anticipation of an impending event such as hurricane landfall or major tornado outbreaks). These systems add a new dimension to our fixed observational assets by providing additional information where it is needed through targeted observations. Automated Aircraft Reports The ACARS (Aircraft Communications Addressing and Reporting S Â ystem) program was initiated in the 1970s by NOAA and the FAA to put temperature and wind sensors on the shells of commercial aircraft. Current participating airlines include United, American, Delta, Northwest, Southwest, Federal Express, and the UPS. Most ACARS-equipped aircraft provide latitude, longitude, altitude, time, temperature, wind direction, and wind speed. Some aircraft also provide turbulence data in the form of either vertical acceleration or eddy dissipation rate. A small subset of aircraft, mostly UPS, also provide moisture data. A review of the precision of the data relative to radiosondes can be found in Schwartz and Benjamin (1995). Today, more than 100,000 automated aircraft reports are available over the United States each day from domestic commercial flights, with more than half coming from ascents and descents below 20,000 ft. The
OBSERVING SYSTEMS AND TECHNOLOGIES 109 main source of reports is the Aircraft Meteorological Data Relay (AMDAR) system, which collects wind and temperature reports from nearly 1500 U.S aircraft operated by major long haul carriers. Moisture measurements aboard a few aircraft (Water Vapor Sensing System, version 2âWVSS2) are being tested. A program called TAMDAR (for Tropospheric AMDAR), started in 2003, collects information from short-hop carriers that generally fly in the mid-troposphere, below the flight levels of long haul carriers. The T Â AMDAR sensor is a lightweight (1.5 lb), low-drag (0.4 lb @ 200 knots), low-power device designed for easy installation and retrofit to any aircraft. The TAMDAR system measures temperature, relative humidity, winds, icing, turbulence, and position using GPS during ascents and descents (Moninger et al., 2008). When operating under typical high-resolution settings, ascent and descent observations are made at 10-hPa (100-m or 300-ft) pressure intervals up to 200 hPA (6000 ft) above ground level. Observations higher than 200 hPa above ground level are made at 25-hPa intervals. If an observation has not been made below 20,000 feet (465 hPa) for 3 minutes, then an observation is triggered by time default; if an obser- vation has not been made at heights above 20,000 ft, then an observation is triggered by time default. A program called Measurements of Ozone, Water Vapor, Carbon Monoxide, and Nitrogen Oxides by In-Service Airbus Aircraft (MOZAIC; Marenco et al., 1998; Zbinden et al., 2006) has been used to profile ozone, carbon monoxide, water vapor, and nitrogen oxides upon takeoffs and landings of commercial aircraft in Europe, Asia, and North America. The measurements have allowed determination of seasonal differences in the ozone column, detection of stratospheric air versus tropospheric sources, and some long-term trends in ozone measurements. Focusing on takeoffs and landings in New York City, the authors have determined that New York has a 10 percent higher contribution to tropospheric ozone than do the European or Asian cities studied. Adding other species to commercial aircraft profiles in U.S. airspace, including the lower troposphere, seems to be a promising new technology. Role of Unmanned Aircraft Systems The National Aeronautics and Space Administration (NASA) and NOAA have recently committed to significant development in unmanned aircraft systems (UAS; earlier called unmanned aerial vehicles or UAV). With the advent of these pilotless vehicles, a new class of airborne platforms is now available for targeted observations. A potentially useful adjunct to coastal weather radars, these vehicles hover over hurricanes and help to enable more precise predictions of ground track at landfall. Errors of order
110 OBSERVING WEATHER AND CLIMATE FROM THE GROUND UP 10 km can cost millions of dollars in damage and risk human life if impre- cise decisions are made. UAS have the ability to fly at 60,000-65,000 ft for up to 12 hours, and pilot fatigue is never an issue, no matter how long the mission. Current developments include the introduction of downward pointing radars, downward pointing lidar, imaging cameras in the visible and infrared, and dropsonde technology. Such developments will allow forecasters to better monitor eye-wall replacement, which can be a major factor in hurricane intensification or weakening. These instruments also will have application in forest and brush fires in deriving plume heights and fire spread, fuel moisture characteristics, and smoke dispersion. In a recent UAS demonstration during the California fires of October 2007, NASA was able to advise the FAA and the U.S. Forest Service (USFS) about fire front behavior, which allowed the teams fighting the fires to target activities to control the blazes and to reroute air traffic around areas of hazardous smoke. Finally, in cases of national emergency in urban areas, the ability of UAS to monitor plume dispersion for hours around an event would be hard to match with other observing systems. Targeted Observations Targeted observations are made on demand, as the problem dictates. They are used for both operations and research. Familiar examples of targeted observations are the NOAA and military aircraft missions that target hurricanes and, during the âoffâ season, take measurements in Âwinter storms affecting the western and eastern U.S. coasts. Other examples are aircraft missions to release dropsondes around hurricanes to improve fore- cast accuracy, rapid-scan images programmed for Geostationary Observa- tion Environmental Satellites (GOES; to be discussed later in this chapter), and extra radiosonde releases when severe storms threaten the U.S. main- land. UAS, large and small, are being considered both for targeted and routine observations, especially in remote oceanic and polar regions; and they have potential to âsniffâ out toxic releases (NRC, 2003a). It might also be possible to move a fleet of mobile radar wind profilers into a region threatened by hurricanes. Targeted observations have been used for research since aircraft pen- etrations of thunderstorms during the Thunderstorm Project in the 1940s. Prominent research-based systems include mobile Doppler radars and mesonets (e.g., Weckwerth et al., 2004, Bluestein et al., 2001). Mobile radars include, for example, the C-band SMART radars (Biggerstaff and Guynes, 2000), the X-Pol and X-band Dopplers on Wheels (Wurman, 2001), the University of Massachusetts millimeter radar (Bluestein and Pazmany 2000), and mobile radar wind profilers (e.g., the Mobile GPS/
OBSERVING SYSTEMS AND TECHNOLOGIES 111 Loran Atmospheric Sounding System [MGLASS] operated by the Earth Observing Laboratory [EOL]/NCAR, the Mobile Cross-Chain Loran Atmospheric Sounding System [MCLASS] operated by the National Severe Storms Laboratory [NSSL]). Radiosonde are also released from mobile platforms. Mobile Radiometers supplied by the Desert Research Institute were deployed during the International H2O Experiment (Weckwerth et al., 2004). Mobile mesonets are typically instrumented cars (e.g., Straka et al., 1996), or rapidly-deployable instrumented towers, such as the Texas Tech University Stick-Net, which can be deployed in the time it takes to drive to the phenomenon of interest (installation time less than 3 minutes). Rapidly deployable surface stations are also available for supplemental observations during wildfires (mobile Remote Automated Weather Stations, RAWS) and many other emergency management and public safety applications, such as deployment after September 11, 2001 in New York City and during major sporting events such as the Super Bowl. Airborne remote sensing has become important in operations and research. The P3 aircraft used to penetrate hurricanes are equipped with a horizontally scanning C-band radar and a vertically scanning X-band D Â oppler radar. A Navy P3, operated jointly with NCAR, also has an air- borne Doppler radar. Both have been used extensively in field programs to investigate convective precipitation, including mountain precipitation, storm initiation, and hurricane landfall. Airborne Doppler lidars and dif- ferential absorption lidars have also been flown. Surface-Based Network Collaborations The desire to use data from multiple networks has led in recent years to âcollectives,â which combine the data from a number of networks, offering easier access to shared data and quality checking for the included groups. The two flagship collectives are the Meteorological Assimilation Data Ingest System (MADIS; Miller et al., 2005), which is based at ESRL/NOAA and is essentially NOAAâs attempt to create a network of networks, and MesoW- est (Horel et al., 2002), which is based at the University of Utah and is the primary source of surface mesonet data for many end users across the country. NOAAâs Hydrological Automated Data System (HADS) provides real- time data from more than 13,000 river and weather sites. Another impor- tant collective is NorthwestNet, based at the University of Washington. More recently, the Federal Highway Administration has started to develop â See http://www.atmo.ttu.edu/TTUHRT/WEMITE/sticknet.htm. â See http://www.nws.noaa.gov/oh/hads/.
112 OBSERVING WEATHER AND CLIMATE FROM THE GROUND UP the Clarus (Latin for âclearâ) System to integrate surface transportation weather observations over the United States, and data from many state departments of transportation have resided on MADIS for years. The Consortium of Universities Allied for Hydrological Sciences has been devel- oping a Hydrologic Data Access System that provides access to data not only from networks (e.g., Ameriflux and Long-Term Ecological Research Network [LTER] sites) but also from North American Regional Reanalysis. RAWS are operated by multiple agencies for air quality and fire weather applications. AIRNOW links air quality data from multiple locations for air-pollution applications. On a smaller scale, the state climatologists in some states have begun efforts to form a collective among the networks within their states. (e.g., South Carolina and Iowa). Finding: Surface network collaborations represent a significant step for- ward and serve to achieve improved quality-checking, more complete metadata, increased access to observations, and broader usage of data serving more than one locally driven need. SPACE-BASED OBSERVATIONS Satellites are essentially tools for global observations, but part of their utility lies in setting the context for mesoscale surface-based observations. Satellites provide vital information on the evolution of severe storms, and remotely sensed soundings and water vapor imagery contribute to the inter- pretation of the severe storm environment. Satellites also provide informa- tion on surface characteristics such as vegetation and soil moisture, both of which have been shown to be important to storm initiation and the prediction of convective weather. Satellite orbits can be divided into two types, geostationary orbits and low-Earth orbits. In a geostationary orbit, the satellite travels at the same rate as the Earth revolves, affording a constant view from a vantage point 40,000 km above the equator. The distance limits resolution and signal- to-noise ratio but affords a view of evolving storm systems on a routine basis, at intervals of 30 minutes or less. Low-Earth orbiting satellites do not provide a continuous view, but their lower altitude (100s of km above Earthâs surface) enables higher-resolution images and stronger signals, opening up the opportunity to generate many remotely sensed products at intervals of twice a day. A common type of low-Earth orbiting satellite, the sun-Âsynchronous polar-orbiting satellite, samples a point on the Earth at 12-hour intervals. Another low-Earth orbit is the less frequently used pro- â See http://www.cuahsi.org/his. â See http://www.fl.fed.us/rm/pubs/rmrs_grt119.pdf.
OBSERVING SYSTEMS AND TECHNOLOGIES 113 grade orbit (faster than Earthâs angular rotation), which enables sampling through the diurnal cycle over time (e.g., for tropical rainfall in the case of the Tropical Rainfall Measuring Mission [TRMM]), and specialized con- stellations that take advantage of other space-based assets such as GPS for radio-occultation measurements. Satellite instruments use the microwave, infrared, visible, and ultra- violet (UV) parts of the electromagnetic spectrum to probe the atmosphere and the Earthâs surface. Measurements of visual reflectance of sunlight from the surface are the easiest to comprehend for those uninitiated to satellite meteorology. Parameters to be retrieved generally relate to the diminution of sunlight before it reaches the surface (clouds, aerosols, absorbing gases) or reflectance quantities from the surface (albedo, land use, vegetation char- acteristics). Recently, techniques to retrieve aerosol optical depth over the ocean (which has a minimally varying reflectance over a dark surface) has been determined to be possible to about 20 percent precision (Remer et al., 2005). Many gases absorb in the ultraviolet and visible portions of the spec- trum, and this provides the ability to do species-specific sampling. O3, SO2, NO2, CHCO, H2O, and aerosol size information have all been retrieved in the UV and visible spectral channels of space-borne measurements. GOES can view large portions of a single hemisphere. GOES spatial resolution is about 1Ã1 km for visible and 4Ã4 km for infrared data. Measurements are at intervals of the order of ten to a few tens of min- utes, providing the capability to follow cloud evolution. In ârapid-scanâ mode, GOES provides valuable data for mesoscale analysis and nowcasting (Browning, 1982). In addition, cloud or water vapor features can be tracked to estimate winds, providing data between radiosonde or aircraft reports. Precipitation has been estimated from an infrared cloud-top algorithm. NOAAâs next generation geostationary satellite, GOES-R, can deliver skin temperature over land and water but with limited spatial resolution and only in clear sky views, and also can be used to identify cloud top height. GOES-R is planned to have a multi-channel Advanced Baseline Imager that will view the United States at least every 5 minutes with a factor of four higher spatial resolution than the current generation of GOES. The Moderate-resolution Imaging Spectroradiometer (MODIS) on the Terra and Aqua polar-orbiting satellites uses infrared radiation to retrieve vegetation characteristics, which are used as input into the land- surface Â models embedded in experimental numerical weather prediction models. Satellite images such as GOES, the Advanced Very High Resolu- tion Â Radiometer (AVHRR), and MODIS, can deliver skin temperature with limited spatial radiation and only for clear skies. Future images for the National Polar-orbiting Operational Environmental Satellite System (NPOESS) and GOES-R will have greater spatial resolution and thus the potential for more clear fields of view. These satellites also have the âsplit
114 OBSERVING WEATHER AND CLIMATE FROM THE GROUND UP windowâ channels that allow for low-level moisture corrections. The cur- rent GOES series of satellites eliminated the split window capability early in the series and replaced the 12.5-micron channel with one at 13.2 microns to help with cloud height assignment; thus in some areas the current GOES surface temperatures are compromised by moisture. Surface winds are problematic from space, except over the open ocean. Microwave imagers such as the Advanced Microwave Scanning R Â adiometer (AMSR) (Njoku et al., 2003) and the Special Sensor Micro- wave Imager (SSM/I) (Jackson et al, 2001) can give surface water content (standing water and in a very shallow, ~1-cm layer below the surface), but cannot give information on subsurface water. AMSRâs footprint is of the order of 25 km so the resolution in areas of varying topography, where runoff and streamflow may be most important, is problematic. Ground-penetrating radar such as RADARSAT (LeConte et al., 2004) has been used to probe soil moisture, but the instrument does not produce a routine product in this regard. The âDecadal Surveyâ (NRC, 2007a) has identified the L-Band (1-2 GHz, ~20 cm-wavelength) Soil Moisture Active-Passive radar-Âradiometer approach as a promising technology for soil-moisture retrieval and placed it in the top three missions for develop- ment by NASA, but also pointed out that the Hydros mission designed for this purpose was cancelled. Spaceborne radars and lidars have been used to sense precipitation, clouds, and aerosols. Radar satellites such as RADARSAT, TRMM and CloudSat are in orbit and are returning information with high vertical reso- lution from these pulsed measurements. In the visible regime, the Cloud- Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO) mission is returning very high vertical resolution images of clouds and aerosols. These active sensors promise to be able to retrieve data well down into the lower troposphere with high resolution, which is of value to the mesoscale observations of this study. TRMM has demonstrated the use of space-based radar measurements to estimate rainfall rates over the ocean while also providing profiles of pre- cipitation and observations of lightning using a high-speed, charge-coupled device detection array. TRMMâs 13.8-GHz precipitation radar electroni- cally scans a 280-km-wide swath with a spatial resolution of 4 km. Scien- tists at NASAâs Goddard Space Flight Center used TRMM data as a baseline to calibrate polar-orbiting passive microwave sensors and geoÂstationary infrared sensors to produce 3-hourly global precipitation maps. The lidar on CALIPSO is a two-wavelength (532 and 1064 nm) p Â olarization-sensitive lidar that provides 30-60 m vertical resolution pro- files of aerosols and clouds within its 100-m field of view. CALIPSO was â See http://trmm.gsfc.nasa.gov.
OBSERVING SYSTEMS AND TECHNOLOGIES 115 launched in tandem with CloudSat, which carries a 94-GHz nadir-looking radar Cloud Profiling Radar (CPR) that measures the power backscattered by clouds as a function of distance from the radar. The CPR profiles clouds along the satellites ground track with a horizontal resolution of 2 km while providing information with 500-m vertical resolution on cloud water and ice concentrations, cloud thickness and cloud base and top height. CloudSat and CALIPSO are flying in a formation called the âA-Trainâ with Aqua, PARASOL,10 and Aura. While satellites are mesoscale-resolving in the s Â patial domain, the infrequent time domain sampling suits them for climate statistics but severely limits their utility either for monitoring or predicting mesoscale events. Space-Based Soundings Satellites provide atmospheric soundings using both infrared and microwave remote sensing. The NOAA High Resolution Infrared Radia- tion Sounder (HIRS) and the GOES sounder provide vertical profiles of temperature and moisture, as well as other variables. The NOAA polar- orbiting satellites that fly Advanced Microwave Sounding Units (AMSU-A and AMSU-B) provide similar thermodynamic information. In both cases, sampling in the spectral region at the center of the absorption band yields radiation from the upper levels of the atmosphere (i.e., radiation from below has already been absorbed). Radiation signals at wavelengths increasingly distant from the center of the absorption bands are from successively lower levels of the atmosphere. This smears out the temperature and moisture, particularly in the lower atmosphere, limiting but not eliminating their u Â tility at the mesoscale. Moreover, infrared soundings require clear skies. More vertical resolution can be obtained using more portions of the spectrum. Hyperspectral infrared sounders, such as the Atmospheric Infra- red Sounder (AIRS) on NASAâs Aqua satellite, uses thousands of spectral bands in the infrared spectrum with greater accuracy and vertical resolution than before, although still not of radiosonde accuracy or vertical resolution, particularly in the lower atmosphere. The EUMETSATâs11 Infrared Atmo- spheric Sounding Interferometer instrument measures atmospheric trace gases in over 8000 channels. The exploitation of such data for mesoscale applications is being investigated. Earlier, we discussed the analysis of signals from GPS satellites to infer the amount of water vapor in a vertical column of air. The technique of radio occultation (RO) greatly expands the utility of GPS; it results in 10â Polarization & Anisotropy of Reflectances for Atmospheric Sciences coupled with Obser- vations from a Lidar. 11â Network of European Meteorological Services.
116 OBSERVING WEATHER AND CLIMATE FROM THE GROUND UP measurements of electron density in the ionosphere, temperature soundings in the stratosphere, and temperature and moisture soundings in the tropo- sphere. Radio-occultation measurements from the Taiwan-U.S. ÂCOSMIC/ FORMOSAT-3 mission12 (Anthes et al., 2008) are providing real-time tem- perature soundings with roughly 500-m vertical resolution in all kinds of weather, since the radio waves are not affected by clouds or precipitation. However, the horizontal resolution is 150-200 km (Ware et al., 1996), a drawback for mesoscale applications. From an altitude of 20,000 km, the GPS satellite sees the low-Earth orbiting satelliteâs rising or setting over the Earthâs surface, hence the term âoccultation.â The speed of the radio waves between the two satellites is a function of the atmosphereâs radio refractive index. Virtual temperature profiles can be calculated from a number of satellite-to-satellite paths, with excellent accuracy from the mid-troposphere upward, where there is little moisture. Sophisticated assimilation techniques that combine RO data with information from prediction models have been able to extract moisture information from the RO data in the lower troposphere, useful for global NWP. Satellite-based sounding information of other quantities, such CO2, CO, O3, and CH4, are of increasingly high quality, but they donât fully meet the requirements for this study, since many instruments derive only full- c Â olumn quantities (column O3, for example), and it is difficult, if not impos- sible, to untangle the planetary boundary-layer information from those profiles. Many sensors, designed to give many levels of vertical resolution, have not lived up to billing and give only one or two pieces of independent information, which tend to peak higher in the troposphere than would be useful for planetary boundary-layer application. Ultraviolet and infrared instruments may get such a large portion of their orbit level radiances from high in the atmosphere that they cannot even see the surface. And finally, clouds as an obscurant are a major limitation in making routine surface observations since approximately 70 percent of the pixels are contaminated by cloud on average. OBSERVATIONAL CHALLENGES The Surface Challenge While much of the technology involved in surface meteorological mea- surements is reasonably mature, important challenges remain. Land-surface properties, especially soil moisture, are only measured in scattered areas, 12â Constellation Observing System for Meteorology, Ionosphere, and Climate/Formosa SATellite.
OBSERVING SYSTEMS AND TECHNOLOGIES 117 yet this has been identified as an important variable in numerical weather prediction, and for many agricultural applications. Likewise, measurements of precipitation type and amount, especially for frozen precipitation in real time, are important for aviation (de-icing aircraft, keeping airports open) and road transportation (informing decisions by road managers regarding plowing and road-treatment chemical application). The Challenges of Geography and Urbanization Although maps show an impressive number of meteorological obser- vations, zooming in inevitably reveals extensive gaps in relation to known mesoscale variability. This is particularly true for soil-temperature, soil- moisture, and air-pollution measurements, as just mentioned. In addition, there are regional-scale surface station deficiencies for real-time reporting of standard meteorological data. While mesoscale and convective-scale phenomena can occur anywhere in the United States, it is not necessary to measure all atmospheric vari- ables at sub-kilometer scales at all locations in order to produce accurate and useful analyses and forecasts. However, there are three regions (urban areas, mountains, and coastal zones) for which nature and/or people have created structures of significance on such small spatial scales that special measurement and network strategies are needed. These structures can cre- ate very strong gradients in atmospheric (and chemical) variables across short distances that are of vital importance to life and property. Whereas measurements over homogeneous terrain are intrinsically representative of a broader area, data in small-scale three-dimensional environments are often representative of only a tiny volume. Moreover, urban areas, mountains, and coastal zones all have special needs. All three create their own weather, which is often poorly resolved in synoptic NWP models. Considering the importance of water storage in the snowpack and reservoirs and hydroelectric power generation, and the danger of traveling in the winter or fighting forest fires in the summer, the needs for observations in the mountains go beyond those for weather fore- casting alone. Coastlines and cities, both of which have high concentrations of people, also take on special importance, particularly when one considers the need for observations to respond to a release of toxic substances, to treat the roads in response to an ice storm or blizzard, or to evacuate people in advance of hurricane landfall. Urban Areas High-resolution weather information in urban areas is vital because of the greater population density coupled with the added complexity introduced
118 OBSERVING WEATHER AND CLIMATE FROM THE GROUND UP by large buildings (and possibly terrain and coastal features as well). The impacts of typical weather phenomena are magnified in cities; for example, heavy rains can cause severe flooding, snow and freezing rain can disrupt transportation, and severe storms and accompanying lightning and high winds can cause power failures. Urban dwellers also are more susceptible to public health and safety issues such as heat stroke, heavy air pollution, and terrorism. Large urban areas also impact weather and atmospheric struc- ture in various ways. Urban heat islands result from the combined effects of changed thermal and radiative properties of the surface, anthropogenic emissions of sensible heat, and changes in the exchange of water and the corresponding impact on the radiation budget. Changes in surface rough- ness in urban areas also affect the exchange of heat, mass, and momentum between the surface and the atmosphere, as well as the depth of the urban mixed layer. Hydrological processes also are altered to a significant degree as a result of buildings and pavement that affect runoff and stream flow. Large urban areas may influence the genesis, intensity, and movement of convective storms and frontal boundaries. There are unique issues related to air quality and terrorism, which were discussed in Chapter 3. There is a pressing need to improve our ability to characterize and forecast urban weather. The increasing spatial resolution of NWP models allows for the opportunity to address urban meteorology and impacts to a greater degree. However, as pointed out by the 10th Prospectus Develop- ment Team of the U.S. Weather Research Program (Dabbert et al., 2000), improving short-term predictions of weather and air quality in urban fore- cast zones requires improvements in our measurement and modeling capa- bilities. Improving our capabilities requires special consideration of the urban environment and the âurbanizationâ of our meteorological measure- ment and modeling components. Recent studies are showing that improve- ments in NWP and air quality dispersion require better descriptions of urban surface fluxes and the vertical structure of the urban boundary layer (Baklanov et al., 2006). These improvements present new challenges for urban observing systems, which need to characterize flows and constrain models operating at scales of a few hundred meters. The measurement challenges include the following (Baklanov et al., 2008): (1) Availability of suitable and representative instrument sites, allowing for security, power, data transmissions, neighborhood convenience, public safety, accessibility, and planning permission; (2) Height and positioning of sensors to meet the needs of adequate reference height, so that the appropriate surface type is within the upwind fetch and observational footprint for sensors; and (3) Sufficient number of sensors to be deployed within the city area as well as at a number of reference rural sites so that influences due to the city can be differentiated from the day-to-day and diurnal changes under various prevailing meteorological situations.
OBSERVING SYSTEMS AND TECHNOLOGIES 119 EScan Panels FIGURE 4.11â Conceptual design of microwave radar antenna panels mounted on 4-12.eps the corner of a building (D. McLaughlin, UMass-Amherst/CASA). bitmap image with vector arrows & type These measurement issues are being studied in testbeds such as the Helsinki study discussed earlier, the Pentagon Shield program (Warner et al., 2007), and in urban field experiments.13 Lessons learned here and elsewhere need to be factored into the urban component of the national mesoscale observing network. Urban networks provide unique challenges such as the need for three-dimensional measurements at dense scales and communications. In addition, the sensors cannot be deployed easily, and dealing with building architectural codes, real estate costs, and societal acceptance becomes very important. These challenges are already being addressed in other disciplines such as cell phone antenna deployments and in weather via a planned CASA project. Figure 4.11 shows a conceptual deployment scenario of low-cost microwave radar sensors in an urban environment where the radar antenna panels are attached to the edges of the taller buildings. The electronic-scanning sensors merge seamlessly with the background and have no moving parts (McLaughlin et al., 2007). Also shown in the figure are the communication antennas. 13â E.g., the December 2007 special issue of Journal of Applied Meteorology and Climatology on Joint Urban 2003.
120 OBSERVING WEATHER AND CLIMATE FROM THE GROUND UP Mountains Mountains affect the weather by initiating convection and deep snow- falls, focusing water or wind into narrow valleys, and generating turbulence aloft and severe (â¥100 mph) windstorms at the surface on their lee side. They present the danger of slick roads, high winds, poor visibility, and avalanches, rockfalls, and mudslides. They also cause significant down- stream effects in weather and stream flow, the latter causing water resource management challenges up to 1000 km. Knowing the water content in the mountain snowpack is important to water managers and their customers. At the same time, mountains present special observational challenges. The weather conditionsâtemperature, winds, and precipitationâaround moun- tains are so variable that the small number of measurement sites cannot capture the complexity. Radars and data transmission are both limited by blockage, and traditional flux measurements in complex terrain are difficult to interpret correctly. Measurement sites are difficult to install and main- tain. Forest fires present a particular challenge, because wind and moisture measurements in remote terrain are critical. Current observations in the mountains can be characterized as sam- pling sparse but sampling smart. Decades of experience have determined where Snowpack Telemetry measurements are sited, and methods are being developed to incorporate satellite information. Likewise, state DOTs know the weather-vulnerable portions of major roadways and the locations of site stations. The dangerous stretches of major roads are often equipped with web-cams to help travelers. Larger metropolitan areas have instrumented the watershed upstream to alert them of the possibility of flash floods. At the mesoscale and smaller scales, challenges remainâparticularly with respect to convective precipitation and wildfires. Because mountains block radar beams, many areas are without coverage. We have the tools to begin to address this, for example, gap-filling radars and lidars that operate in adaptive-collaborative modes with rain gauges, stream gauges, and satellites. As cell-phone towers proliferate, these offer platforms not only for radars and lidars but also for communication of data from remote sites. High-resolution numerical models need to be part of the observational mix. Mountains provide strong forcing, making precipitation and wind pat- terns more predictable. The combination of good upstream conditions with some boundary-layer, surface, and radar data that the model can assimilate has the potential to provide the three-dimensional picture needed by fire meteorologists, snowpack and runoff analysts, and flash-flood or downslope windstorm forecasters. Thus the components needed for a âMountain Netâ are included in our architecture as proposed, the primary challenge being to address the severe under-sampling problem.
OBSERVING SYSTEMS AND TECHNOLOGIES 121 The Coastal Zone Coastal regions have both natural and made-made features that create complex spatial and temporal variability in weather and sea-state condi- tions, much of which may go undetected. For example, prevailing offshore flow could be replaced by a sea breeze along one stretch of coast while adjacent stretches remain offshore, affecting forecasts of convective initia- tion and energy demand. Varying winds also can affect the destination of a hazardous chemical leak and the towing of ships and barges in a harbor. Coastal fronts can move onshore ahead of winter storms, affecting where different types of precipitation fall (solid, partly frozen, and unfrozen). Unmeasured air-sea interactions occur offshore, creating moisture and sta- bility conditions that are ripe for severe weather outbreaks in the return flow regions. The vulnerability of coastal zones is increasing annually, as more coastal regions become large population centers. Coastal counties are growing three times faster than other U.S. counties, and coastal and marine waters are an annual tourist destination for 90 million Americans. In addition, many coastal regions have significant topography, suggesting that their special observing and network needs are congruent with those of the Urban Net and Mountain Net described above. Additional requirements, though, exist for offshore weather and sea-state data, including profiling of winds, temperature and moisture above the surface, and temperature, current, and salinity at and below the surface. Thus the U.S. mesoscale network of networks should include a suite of additional buoys and land stations, and remote sensing capabilities extending 100-200 km from the coast. The Planetary Boundary Layer Challenge One of the most difficult to measure and yet one of the most important parameters is the height of the daytime and nighttime planetary boundary layer (PBL). Driven in the daytime by heating of the surface and convec- tion and driven at night by winds and infrared radiative cooling of the surface, the PBL height is critically important in forecasting constituent concentrations in numerical models (since this is the height of the box into which constituents mix and react). It is now believed that the imprecision with which the PBL height is known is a major source of uncertainty in the predictive capability of current numerical chemical forecast models. It really is astounding after nearly sixty years of remote sensing observations in meteorology that such an important meteorological variable is not mea- sured with regularity throughout its diurnal cycle. The only area of relative strength relates to winds from ultra-high fre- quency (UHF) and very high frequency (VHF) wind profiles when combined
122 OBSERVING WEATHER AND CLIMATE FROM THE GROUND UP with AMDAR and TAMDAR observations. This combination thoroughly captures the synoptic scale, and also some of the larger mesoscale circula- tions. However, the characteristic spacing of radar wind profilers is too large, often missing medium-sized mesoscale circulations that spawn dis- ruptive and severe weather. The number of commercial airline observations has a large-amplitude diurnal cycle, which deprives the composite observing system of needed data for ~8 hours per day and leaves the system vulnerable during large storms (or terrorist attacks) when there are wholesale flight cancellations. The national condition for thermodynamic, trace-gas, and aerosol profiling is one of significant inadequacy to address mesoscale predic- tion needs. A major improvement in thermodynamic profiling is needed. Radiosonde sites are several hundred kilometers apart and only address the synoptic scale. The vertically resolved water vapor field, especially in the BOX 4.2 An Example Core Observing Site to Address the Planetary Boundary Layer Challenge For the last 20 years, Howard University of Washington, D.C., has operated a research station at Beltsville, Maryland. Since 2001, when a NOAA Center for A Â tmospheric Sciences was founded at Howard as part of a NOAA Cooperative Agreement, the Beltsville facility has grown into a high-level core mesoscale observation site. A tall tower to measure CO2 fluxes was installed by the UniverÂ sity of Virginia, and a Â Raman lidar was constructed in cooperation with NASA. Radiosonde observations and ozonesonde releases have been carried out for validation of NASAâs Tropospheric Emission Spectrometer and Ozone Monitoring Instrument. NOAA has contributed many of the radiosondes as part of its modernÂ ization program and the Pennsylvania State University has contributed the ozone soundings as part of the NASA INTEX Ozonesonde Network Study (IONS). Baron Meteorological Services has contributed a weather radar to the site. The EPA and the State of Maryland have contributed a radar wind profiler and a ground-based chemical monitoring capability (PM, O3, NOx) to the site. The U.S. Department of Agriculture (USDA) has contributed a shadowband radiometer to the site to meaÂ sure aerosol optical depth, and NASA has contributed an AERONET site. Surface energy fluxes and subsurface temperature and moisture are measured at the site. Sonic anemometers measure turbulence at the site. Surface solar radiation fluxes (as in the Baseline Surface Radiation Network, NOAA) are Âbeing routinely made at the site. Arguably, Beltsville is the type of station which can be expected to arise from the efforts recommended in this report. Multiple agencies, with disparate needs, can contribute to a single site and leverage resources. It is interesting that this site was founded and is operated by a Minority Serving Institution, which clearly
OBSERVING SYSTEMS AND TECHNOLOGIES 123 lowest 1 km, is most critical, being essential for improved prediction of all high-impact weather. The needs of chemical weather predication are on a similar plane, requiring national-scale coverage of major pollutant species including aerosols, thereby enabling urban and regional pollutant forecasts. Some research stations have been established that include many of the core, ground-based remote sensing systems that supply these types of observa- tions (see Box 4.2). Yet, there is no national coverage of sufficient scope to address the planetary boundary-layer challenge. Recommendation: As a high infrastructure priority, federal agencies and their partners should deploy lidars and radio frequency profilers nationally at approximately 400 sites to continually monitor lower trospospheric conditions. could not have accomplished a project of this scope without contributions from the federal and private sector. FIGURE 4.2.1â The Howard University Beltsville site showing the instrumentation component and training. SOURCE: Whiteman et al. (2006). Figure 4-13 now Figure 4.2.1
124 OBSERVING WEATHER AND CLIMATE FROM THE GROUND UP Wind, diurnal boundary-layer structure, and water vapor profiles are the highest priority for a network, the sites for which should have a char- acteristic spacing of approximately 150 km but could vary between 50 and 200 km based on regional considerations such as those just discussed for urban areas, mountains, and coastal zones. Such observations, while not mesoscale resolving, are essential to improved performance by high- r Â esolution numerical weather prediction models and chemical weather pre- diction at the mesoscale. Through advanced data assimilation techniques, data from these 400 sites, when used in combination with geostationary satellite measurements, GPS constellation âwet delayâ measurements, and commercial aviation soundings, could effectively fill many of the critical gaps in the national observing system. Any sensors that measure air chemistry or aerosol properties above ground should be located with or near the meteorological profilers. Most chemical models use vertical grid spacing of 100-200 m in the lowest 3 km (finer near the surface), but chemical and aerosol measurements at even two or three levels within the lowest 3 km would be a marked improve- ment over present capability. Above 3 km, satellite measurements become increasingly effective at greater altitudes. Some of the measurements could be from towers, others from remote sensing by lidars or differential optical absorption spectroscopy.14 Challenges for Space-Based Observations The âDecadal Surveyâ (NRC, 2007a) has identified a path forward for the next generation Earth observational satellite system for the United States. Measurements that are relevant to mesoscale applications include soil moisture using the L-band, soil composition and vegetation character- ization from a hyperspectral spectrometer, columns of atmospheric trace gases to high horizontal resolution, aerosol and cloud profiles, land-surface topography, temperature and humidity soundings, tropospheric winds that donât depend on feature tracking (from Doppler lidar), and subsurface water. The United States was expected to play a leading role in developing many of the space-based capabilities mentioned above, which in turn would have contributed substantially to mesoscale observations of the Earth, its ocean, and atmosphere. However, as pointed out in the âDecadal Surveyâsâ preliminary report (NRC, 2005), âThe national system of environmental satellites is at risk of collapse.â Further deterioration in the U.S. plans led to an even more pessimistic assessment in the final report (NRC, 2007a): âThose concerns have greatly increased in the period since the interim 14â Described at http://www.atmos.ucla.edu/~jochen/research/doas/DOAS.html.
OBSERVING SYSTEMS AND TECHNOLOGIES 125 report was issued, because NASA had cancelled additional missions, and NOAAâs polar and geostationary satellite programs have suffered major declines in planned capabilities.â The âDecadal Surveyâ links observations in solid earth, water, weather, climate, health, and ecosystem science areas to meeting societal challenges regarding water, food, and energy security, early warnings of hazardous weather, ecosystems services, and improvements in public health and envi- ronmental quality. Specific recommendations are made to both NOAA and NASA concerning GOES-R hyperspectral sounding capability, the elimina- tion of climate monitoring sensors in NPOESS, deletion of the Conical image: Rain rate Scanning Microwave Imager/Sounder, and removal of key meteorological sensors from the early-morning orbiting satellite (0530 LST Equator crossing), and a series of other missions relevant to this study. From a mesoscale perspective, the most disturbing finding was the elimination of hyperspectral infrared temperature and water vapor sounding capability from geostationary altitude. Support for the âDecadal Surveyâ conclusions is widespread. The National Weather Association, primarily representing operational forecasters, strongly advocates for âinclusion of a capable high spectral resolution atmospheric infrared sounder on the next generation of GOES-R series of spacecraft.â The American Meteorological Societyâs Committee on Satellite Meteorology and Oceanography issued a consensus statement âOn the Importance of Deploying a GEO Advanced Sounder without Delay.â Further, a recent NRC workshop on âEnsuring the Climate Measurements from NPOESS and GOES-Râ found strong advocacy for geostationary hyperspectral sounding, as did the coincident WMO âWorkshop on the Re-design and Optimization of the Space-based Global Observing System.â In conclusion, satellites will play an increasingly important role in mesoscale observation, but limitations of frequency, resolution, and preci- sion near the surface mean that satellite profiles will not replace ground- based observations in the near future. Finding: It is a national imperative to sustain and improve operational geostationary satellite observations as a critical adjunct to the surface- based mesoscale network. Observations from geostationary orbit are unique and inherently Mesoscale, owing to the high rate of time domain sampling and excellent horizontal resolution. Visible and infrared imagery are invaluable to severe weather forecasts and warnings. Estimates of assimilation of radiances, cloud-drift winds, and free troposphere water vapor enable the initializa- tion of global and mesoscale models. Over land, the vertical resolution of water vapor and temperature data normally obtained from geostationary
126 OBSERVING WEATHER AND CLIMATE FROM THE GROUND UP orbit is not independently sufficient but absolutely essential. Continuous cloud layers prevent infrared soundings below cloud top; however, micro- wave imaging array technology (Lambrigtsen et al., 2006) offers a useful lower-resolution alternative under cloud cover. On the other hand, it is impractical to establish a stand-alone surface-based network with adequate horizontal resolution throughout the depth of atmospheric boundary layer and lower troposphere. Soundings obtained from ground-based profilers (including aircraft of opportunity) and geostationary satellites complement each other optimally, each of their strengths compensating for the otherâs relative weaknesses. Recommendation: As a high satellite instrument priority, NASA and NOAA, in cooperation with foreign space agencies, should seek to improve the quality of geostationary satellite water vapor and tempera- ture soundings within continental atmospheric boundary layers. Infrared hyperspectral soundings and soundings from microwave syn- thetic thinned aperture arrays, each in geostationary orbit, offer unique opportunities to improve mesoscale prediction. While potentially costly, the benefits from improved geostationary soundings would be large, likely enabling more skillful forecasts of convective rainfall and attendant severe weather and flooding. The geostationary platform is unique among satel- lites, offering the sampling frequency required in this application. GLOBAL CONTEXT AND INFRASTRUCTURE Much of the data collected are managed globally through the evolving Global Observing System (GOS),15 which is coordinated by the WMOâs World Weather Watch. Data from the the GOS are used for a variety of applications that span time scales from nowcasting to climate, and include land, ocean, atmosphere, and ecological applications. The GOS provides valuable examples of how a variety of user needs and requirements for various applications are addressed as well as how important areas such as data exchange are handled. GOS is composed of two major subsystems, space- and ground-based. Each may be thought of as a system of systems. The ground-based subÂ system provides observations from surface observing stations on land, upper air observing stations, ships at sea, moored and drifting buoys, and aircraft. While some of these systems are owned and operated by WMO members, the aircraft system is operated by various airlines and coordinated 15â Detailed information on the observing system component of the GOS can be found at http://www.wmo.int/pages/prog/www/OSY/gos-components.html.
OBSERVING SYSTEMS AND TECHNOLOGIES 127 within WMO through the AMDAR System Panel.16 Some of the observing systems are coordinated with other international organizations (mainly the Global Ocean Observing System [GOOS], Global Terrestrial Observing System [GTOS], and Global Climate Observing System [GCOS]). Data from the space-based subsystem of the GOS17 are provided by operational satellites in low-Earth and geostationary orbits and selected research satel- lites in low-Earth orbits. Those satellites are operated by various countries or consortia of countries with WMO activities through mechanisms such as Coordination Group for Meteorological Satellites (CGMS) and Committee of Earth Observing Satellites (CEOS). The GOS Ground-Based Sub-System Over land, a relatively sparse network of nearly 11,000 stations Âdelivers observations of conventional meteorological parameters. About 4000 of those stations comprise the Regional Basic Synoptic Networks, whose data are exchanged globallyÂ in real time in compliance with WMO Regulation 40.18 Over the oceans, ships and moored and drifting buoys also pro- vide information for GOS. On any given day, about 2800 ships and 900 drifting buoys provide near-surface meteorological parameters as well as sea-surface temperature.19 Solar radiation observations, surface lightning network observations, and tide-gauge measurements are also provided via the GOS, but in limited numbers. Upper air observations are provided mainly by land-based radiosonde and aircraft data, with a limited number of observations from ground-based wind profilers and radiosonde releases from ships at sea. Close to 900 land-based upper air stations provide radiosonde sound- ings to the GOS twice a day: at 1200 and 0000 UTC. The AMDAR system provides observations of temperature and wind from commercial aircraft at flight level as well as soundings during ascent and descent. As noted by the 2007 WMO Expert Team on the Evolution of the GOS,20 the global AMDAR program exchanges between 220,000 and 250,000 observations 16â âThe goal of the Panel shall be to enhance the upper-air component of the Observing System of the World Weather Watch through cooperation among Members in the acquisition, exchange and quality control of meteorological observations from aircraft using automated reporting systems.â The AMDR Panelâs goals are found at http://www.wmo.int/amdar/Goal_ TOR.html. 17â Detailed information about the space-based component of the GOS can be obtained from the WMO Space Program web site: http://www.wmo.int/pages/prog/sat/index_en.html. 18â Regulation 40 addresses the free exchange over the Global Telecommunications System of 6-hourly RBSN and all upper air, ocean, and satellite data (some Members provide surface observations on an hourly basis). 19â See http://www.wmo.int/pages/prog/www/OSY/gos-components.html. 20â See http://www.wmo.int/pages/prog/www/OSY/Reports/ET-EGOS-3_Final-Report.pdf.
128 OBSERVING WEATHER AND CLIMATE FROM THE GROUND UP per day over the WMO Global Telecommunications System. Most AMDAR observations are in the Northern Hemisphere, and programs like EUCOS (EUMETNET Composite Observing System) are working to optimize AMDAR ascent and descent data for use by EUCOS member countries. For example in 2006, EUMETNET-AMDAR provided approximately 750 soundings per day.21 The RBSN observing stations and conventional upper air network do not all report on a routine basis, with the performance varying greatly by WMO region.22 Reports from stations over the United States are very reliable. In addition to the GOS, specialized observing networks such as the Global Atmospheric Watch for chemistry and the World Hydrological Cycle Observing System provide data that may or may not be in real time. Approximately one-fourth of the RBSN stations make up the Global Cli- mate Observing System (GCOS) Surface Network,23 and approximately 20 percent of the upper air sites make up the GCOS Upper Air Network. As with the GOS, performance of the GCOS sub-set of the GOS is not at 100 percent. The GOS Space-Based Subsystem The space-based subsystem of the GOS embraces the concept of a com- posite observing system with research and operational satellite data used in synergy.24 Data are provided by both operational satellites and low-Earth orbit research satellites. Examples of the research products are hyper- spectral sounding data from AIRS, altimetry measurements from JASON, precipitation measurements from TRMM, and sea-surface winds from ENVISAT. Much of the satellite data flowing into the GOS are used for routine analysis, nowcasting, and forecasting applications at the National Meteorological and Hydrological Services (NMHS) across the globe. Global NWP centers use the data for a variety of forecast guidance products. How the GOS is expected to evolve over the coming decades was recently discussed in WMO Technical Document No. 1267, âImplemen- tation Plan for Evolution of Space and Ground-based Subsystems of the 21â See http://www.wmo.ch/pages/prog/www/OSY/Meetings/ET-EGOS_Geneva2006/Doc4-5. doc. 22â GOS performance is routinely monitored by major NWP centers (see for example http:// www.ecmwf.int/products/forecasts/d/charts/monitoring/coverage/), however, the WMO for- mally evaluates GOS performance during special observing periods each year and the per- formance for various regions can be accessed from the reports of the Commission on Basic Systems reports on the following web site: http://www.wmo.int/pages/prog/www/CBS-ÂReports/ CBSsession-index.html. 23â See http://www.wmo.int/pages/prog/gcos/documents/GSN_Stations_by_Region.pdf. 24â See http://www.wmo.int/pages/prog/www/OSY/gos-components.html.
OBSERVING SYSTEMS AND TECHNOLOGIES 129 GOS,â25 which makes specific recommendations concerning the evolution of the space-based and surface-based subsystem of the GOS. Those recom- mendations were based on guidance from the Rolling Requirements Review process26 as well as observing system experiments, and observing system simulation experiments performed by various NWP centers. Results from these experiments are presented at WMO-sponsored workshops, such as the Fourth WMO Workshop on the Impact of Various Observing systems on NWP. Because of the long lead times for satellite systems, plans for the evolution of the space-based portion of the GOS have been based mainly on the long-term planning of both operational and research satellite operators. Future research missions will continue to contribute to the space-based component of the GOS while influencing its evolution. Those planned research missions include investigations of atmo- spheric chemistry and trace gases, the Earthâs gravity field, soil moisture and ocean salinity, atmospheric winds using lidar, disaster and environmental monitoring, integrated atmospheric column water vapor, cloud ice content, cloud droplet properties and distribution, aerosols, and polar ice and snow water equivalent. Instrumentation under development to accomplish these measurements include space-borne lidar, high-resolution and hyperspectral imaging and sounding instrumentation, active and passive microwave sen- sors, cloud resolving radars, and L-band radars. 25â Information on WMO activities with respect to redesign of the GOS can be found at http://www.wmo.int/pages/prog/www/OSY/GOS-redesign.html, with a link to WMO Technical Document 1267 at http://www.wmo.int/pages/prog/www/OSY/Documentation/ Impl-ÂPlan-GOS_Sept2004.pdf. 26â The Rolling Requirements Review (RRR) process is used to determine how well the GOS is meeting WMO user requirements in a variety of applications area. The RRR procedure consists of four steps: review of user requirements for observations; assessment of the capabili- ties of existing and planned observing systems; critical review (gap analysis), comparing the requirements with system capabilities, in terms of present and planned networks; and state- ment of guidance, which lists conclusions and identifies priorities for action. This information is made available to all users (WMO 2007).