3
Precipitation Data in NOAA Operations

This chapter describes the needs, capabilities, and potential opportunities of the National Oceanic and Atmospheric Administration (NOAA) for using space-based precipitation estimates. To identify the best uses of Global Precipitation Measurement (GPM) mission data at NOAA, the committee first examines NOAA mission requirements for precipitation data and related products. Next, the committee identifies current sources of NOAA operational precipitation data and recommends improvements for these sources in preparation for GPM. Next, five challenges are identified for future space-based precipitation missions. Finally, this chapter outlines the applications of space-based precipitation data in general and the potential applications of GPM mission data specifically.

NOAA MISSION REQUIREMENTS FOR PRECIPITATION DATA AND RELATED PRODUCTS

Global observation of precipitation on a range of time and space scales is essential to achieving NOAA’s mission objectives related to the monitoring and prediction of weather, climate monitoring, many aspects of hydrologic monitoring and prediction, climate data set development, and more (Box 3.1). NOAA maintains or contributes to a wide variety of in situ and satellite-based precipitation measurement systems in support of its mission and the World Meteorological Organization (WMO) World Weather Watch. Because precipitation crosscuts many applications, NOAA’s production, application, dissemination, and archiving activities associated with precipitation data are carried out at multiple centers. NOAA requirements for precipitation data were reviewed at a 2001



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 33
Noaa's Role in Space-Based Global Precipitation Estimation and Application 3 Precipitation Data in NOAA Operations This chapter describes the needs, capabilities, and potential opportunities of the National Oceanic and Atmospheric Administration (NOAA) for using space-based precipitation estimates. To identify the best uses of Global Precipitation Measurement (GPM) mission data at NOAA, the committee first examines NOAA mission requirements for precipitation data and related products. Next, the committee identifies current sources of NOAA operational precipitation data and recommends improvements for these sources in preparation for GPM. Next, five challenges are identified for future space-based precipitation missions. Finally, this chapter outlines the applications of space-based precipitation data in general and the potential applications of GPM mission data specifically. NOAA MISSION REQUIREMENTS FOR PRECIPITATION DATA AND RELATED PRODUCTS Global observation of precipitation on a range of time and space scales is essential to achieving NOAA’s mission objectives related to the monitoring and prediction of weather, climate monitoring, many aspects of hydrologic monitoring and prediction, climate data set development, and more (Box 3.1). NOAA maintains or contributes to a wide variety of in situ and satellite-based precipitation measurement systems in support of its mission and the World Meteorological Organization (WMO) World Weather Watch. Because precipitation crosscuts many applications, NOAA’s production, application, dissemination, and archiving activities associated with precipitation data are carried out at multiple centers. NOAA requirements for precipitation data were reviewed at a 2001

OCR for page 33
Noaa's Role in Space-Based Global Precipitation Estimation and Application BOX 3.1 NOAA’s Five Primary Mission Goals Protect, restore, and manage the use of coastal and ocean resources through an ecosystem approach to management. Understand climate variability and change to enhance society’s ability to plan and respond. Serve society’s need for weather and water information. Support the Nation’s commerce with information for safe, efficient, and environmentally sound transportation. Provide critical support for NOAA’s mission. SOURCE: NOAA, 2004. workshop (NOAA, 2002), and these requirements are discussed in more detail in the following sections. Requirements for Weather Applications The critical requirement for weather-related applications is timely and continuous availability of accurate precipitation data transmitted in WMO formats. NOAA weather-related activities requiring precipitation data include nowcasting (0-3 hours lead time); short-term forecasting (3-12 hours), multiday numerical weather prediction (NWP) forecasts, and preparation and dissemination of centralized forecast guidance. Many of NOAA’s weather-related (and climate-related) operational activities are centered in the National Centers for Environmental Prediction (NCEP). NCEP’s operational needs for global precipitation data include initialization of atmospheric and surface hydrologic (soil moisture) components of coupled NWP models and forecast verification. NCEP’s Environmental Modeling Center needs data describing precipitation characteristics, processes (e.g., phase, type, vertical distribution), and ambient conditions (e.g., temperature, humidity, winds) for improving model physics and data assimilation methodology. NCEP central operations provide services in the form of centralized forecast guidance and analysis products that support the public use of NCEP’s National Weather Service (NWS) forecasts. These products, which include precipitation forecasts, are delivered through the NCEP Hydrometeorological Prediction Center (heavy precipitation forecasts), Storm Prediction Center (severe weather forecasts), Tropical Prediction Center (tropical cyclone forecasts), and Aviation Weather Center. Frequent sampling and timely data availability are also critical for precipitation nowcasts and short-term projections prepared by the National Environmental Satellite, Data, and Information Service (NESDIS) Satellite Analysis Branch.

OCR for page 33
Noaa's Role in Space-Based Global Precipitation Estimation and Application Requirements for Climate Applications NOAA’s climate-related requirements for precipitation data are similar to requirements for weather-related activities with three exceptions: (1) the requirements for timely receipt of precipitation data for operational purposes are generally less stringent, (2) there are more stringent requirements for absolute accuracy, and (3) there is a fundamental need for long, stable, and consistent precipitation time series. NOAA climate-related operational activities require precipitation data for monitoring, diagnosis, and prediction of short-term (seasonal to interannual) climate variability. Precipitation data are also required for climate data set development and for mission-oriented research on climate variability, diagnosis of climate trends, and modeling of climate change. The NCEP Climate Prediction Center provides near-real-time monitoring, assessment, and projections of seasonal-interannual climate variability for use by U.S. agencies with national and international interests, United Nations agencies (Food and Agricultural Organization, WMO), and the public. The Climate Prediction Center’s satellite-based, high-resolution morphing technique (CMORPH) for global precipitation analysis is a key tool for supporting the Climate Prediction Center’s monitoring and diagnostic activities. The NOAA Climate Diagnostic Center requires precipitation data to support its mission of providing diagnostic information on the nature and causes of climate variations, with the goal of predicting these variations. The NOAA Climate Program Office is a focal point for many climate activities within NOAA. The Climate Program Office supports several projects dealing with the development and use of satellite precipitation data sets. These projects include the Climate Change and Detection Project, the Applied Research Center for Data Set Development for transition of Climate Change and Detection Data Projects to NOAA operations, and the Scientific Data Stewardship Program, which governs the production of climate data records. Requirements for Hydrologic Applications Surface hydrology requirements for precipitation data intersect weather and climate needs. They include flash flood forecasts and warnings, monitoring and assessing the impact of drought (e.g., fire risks, crop yields, river stage forecasts), monitoring and predicting runoff from the snow pack in the western United States (which is of paramount importance for water resource management), and other hydrologic information from NOAA’s 13 River Forecast Centers.

OCR for page 33
Noaa's Role in Space-Based Global Precipitation Estimation and Application FIGURE 3.1 Duration of spaceborne, geosynchronous imager missions and their potential extensions (explanations of program name abbreviations are found in the “Visible and Infrared” section later in this chapter). Future missions are subject to change. NOAA’s Ability to Fulfill Its Precipitation Measurement Requirements Increasingly comprehensive and higher-quality satellite precipitation data and data products have become available during the past several years as a consequence of the launch of the Tropical Rainfall Measuring Mission (TRMM) and a number of polar-orbiting satellites carrying passive microwave sensors. Despite these advances, NOAA’s requirements for global precipitation data continue to exceed what is available (NOAA, 2002), and each of the primary sources of space-based precipitation information has different strengths and weaknesses (see Figure 1.2 and Figure 3.1) with respect to fulfilling these data needs. At NOAA’s precipitation workshop in 2001, participants identified NOAA’s requirements that would not be met by existing or planned systems (NOAA, 2002). The deficiencies in space-based observations that were identified can be summarized into three broad categories: data quality and consistency and quantitative description of error characteristics,

OCR for page 33
Noaa's Role in Space-Based Global Precipitation Estimation and Application time and space resolution, and timely availability for operational use. Participants reached the conclusion that to mitigate the deficiencies noted in their report, “substantial improvements in this information are necessary to advance beyond our present capabilities” (NOAA, 2002). Some of the deficiencies have now been mitigated (e.g., higher-resolution global analyses have been developed [in prototype] by combining information from polar-orbiting passive microwave and geosynchronous infrared observations [see Figure 1.2 and Figure 3.1]), but for the most part the deficiencies still exist. The workshop report recognized that many of the existing deficiencies can be significantly mitigated by GPM (see NOAA, 2002, Finding 6 and Recommendation 1). Finding 6 in the workshop report states: “The proposed NASA/Global Precipitation Mission would provide data that would greatly improve NOAA’s ability to monitor and predict weather and climate variability” (NOAA, 2002). Recommendation 1 states: “NOAA should become an active partner with NASA [National Aeronautics and Space Administration] in the Global Precipitation Mission. This system will provide the global three hourly precipitation estimates required by the operational modeling centers. Furthermore, significant improvements in precipitation information for nowcasting, extreme precipitation events, and flash floods will be achieved when geostationary data, gauges, and radars are combined with GPM. Consideration should be given to the establishment of a science team or working group that would define NOAA’s role in and relationship to GPM” (NOAA, 2002). SOURCES OF NOAA OPERATIONAL PRECIPITATION DATA There are two primary sources of operational precipitation data: ground-based observation systems and space-based observation systems. As the GPM core and constellation satellites supplement these sources, they will benefit from being validated against data from a robust ground-based network. This section reviews the status and attributes of ground-based and satellite sources and makes recommendations for improvements that will benefit GPM in particular and global precipitation estimation in general. Ground-Based Sources The continental United States is instrumented with a variety of rain gauges and weather radars that measure precipitation. Both sources have a variety of spatial and temporal sampling approaches that depend on domain and precipitation type.

OCR for page 33
Noaa's Role in Space-Based Global Precipitation Estimation and Application FIGURE 3.2 Distribution of rain gauge data available to NOAA in near real time from three networks: ASOS sites (red), IFLOWS (green), and HADS sites (blue). SOURCE: NWS, 2006a,b,c. Rain Gauge Data Rain gauge data for NWS operations in the continental United States come from multiple agencies in a cooperative network that combines the physical resources of these agencies and is facilitated by good communication links and automation software. The data from these multiple sources undergo quality control and are incorporated in near real time to form an extensive rainfall database. The data are analyzed to map precipitation distribution, determine the potential extent of flooding, and calibrate and validate radar and satellite precipitation estimates. The following observation networks contribute to NOAA operations: (1) the automated surface observing system (ASOS), (2) the Integrated Flood Observing and Warning System (IFLOWS), and (3) the Hydrometeorological Automated Data System (HADS) (Figure 3.2). The HADS data set comes from a number of agencies (Table 3.1). In addition to the three networks mentioned above, data sets from local, state, and federal cooperative efforts, known as “mesonets,” are integrated into near-real-time data streams that feed NOAA operations. As additional sensors are connected into these mesonets through upgraded data links and become accessible on the Internet, further opportunities will emerge from multiagency partnerships that tap into mesonets and observing systems.

OCR for page 33
Noaa's Role in Space-Based Global Precipitation Estimation and Application TABLE 3.1 Rain Gauge Networks in the Hydrometeorological Automated Data System Agency, Data Set Program Number of Gauges Federal, State Wildland Fire 2,400 U.S. Geological Survey 1,836 U.S. Army Corps of Engineers 1,757 NWS 234 Data Collection Service 117 NOTE: HADS data provide near-real-time rainfall accumulation in the continental United States. The distribution of gauges is nonuniform, and the sampling frequency ranges from 1 minute to daily. SOURCE: NWS, 2006b. Whereas rain gauge data are extensive in some regions, nonuniform gauge placement (e.g., Figure 3.2) creates sampling problems such as biases (Sevruk, 1989). In addition, rain gauge measurements have inherent inaccuracies that must be addressed before these data can contribute to the overall precipitation mapping mission (Steiner et al., 1999). Rain gauge data are often treated as “surface truth,” but comparisons with rainfall estimates from radar and satellite rain estimates remain uneven due to these sampling inconsistencies and inaccuracies. Finding: In collaboration with other agencies, NOAA maintains an extensive rain gauge network that provides data in near real time that will contribute to GPM’s calibration and validation efforts. The value of this network to such efforts will be enhanced as data links are upgraded and new mesonets and observing systems become more accessible with rigorous quality control. Recommendation 3.1: NOAA should explore collaborative efforts to augment the existing rain gauge network with additional resources coming online through mesonets that are increasingly used by local, state, and federal agencies to quantify precipitation for many near-real-time applications. NOAA should maintain rigorous quality control and integrate the resultant rain gauge data sets into GPM calibration and validation efforts. Radar Data The NWS operates an extensive network of real-time radar (Weather Surveillance Radar 88 Doppler [WSR-88D] Next Generation Radar [NEXRAD])

OCR for page 33
Noaa's Role in Space-Based Global Precipitation Estimation and Application FIGURE 3.3 Location of NOAA WSR-88D NEXRAD radar sites in the continental United States and the distribution of coverage at altitudes of 4,000, 6,000, and 10,000 feet (tan, orange, and blue circles, respectively). SOURCE: NWS, 2006d.

OCR for page 33
Noaa's Role in Space-Based Global Precipitation Estimation and Application sites (Figure 3.3) around the continental United States, Alaska, Hawaii, Guam, and Puerto Rico that are used in concert with rain gauge data to map precipitation. The NEXRAD network provides near-real-time rain totals (with updates every 6 minutes) that aid in issuing flood watch and warning nowcasts and forecasts. The network’s ability to fully automate the retrieval process and quickly communicate the digital values throughout a region and across the continental United States is of particular value to the NWS River Forecasting Centers. Real-time radar rain estimates are especially good in the eastern two-thirds of the United States, where terrain blockage issues are infrequent (Maddox et al., 2002). Real-time rain measurements permit emergency managers to respond quickly. Combined radar and rain gauge values enable creation of enhanced data sets benefiting multiple user communities (e.g., flood control, agriculture, transportation). NEXRAD data are crucial during rapidly evolving summer thunderstorm events as well as for prolonged and extreme events such as landfalling tropical cyclones. In addition to the positive attributes of the NEXRAD network, it has some shortcomings due to terrain blockage in mountainous areas (Figure 3.3); inability to capture low-level rain because the radar beam rises with distance from the radar site; lack of uniformity of the reflectivity versus rain rate relationship from FIGURE 3.4 Change in effective NEXRAD radar coverage in the western continental United States due to wintertime precipitation and high terrain (right panel) compared with summertime coverage (left panel) when convection and associated higher cloud tops enhance the spatial sampling. SOURCE: Kondragunta, 2005.

OCR for page 33
Noaa's Role in Space-Based Global Precipitation Estimation and Application one location to another and over time; difficulty with mixed-phase and frozen precipitation; decreased wintertime radar coverage in the western United States because low-level precipitation dominates during the winter (Figure 3.4); and data interruptions during extreme events such as hurricanes due to power or communications going offline. The shortcoming related to interruptions can be reduced by upgrading communications systems and building in greater redundancy. The drawback related to accuracy of rain rate estimates can be mitigated by removing biases via TRMM-like precipitation radars (Anagnostou et al., 2001) and by deploying dual-polarization radar (Bringi and Chandrasekar, 2001; Ryzhkov et al., 2005). In addition, some of the shortcomings of ground-based sources in general will be mitigated by higher-fidelity satellite precipitation estimates. Finding: NWS radar precipitation mapping provides critical real-time monitoring and forecasting capabilities that support many NOAA functions and offices. In addition, this data source will be invaluable to GPM calibration and validation efforts. However, the radar network suffers from a number of shortcomings with respect to accuracy and spatial and temporal coverage that can be ameliorated with radar and power upgrades and increased communications redundancy. Recommendation 3.2: NWS should proceed with upgrading the NEXRAD network with dual-polarization radar and should enhance network reliability with upgraded power and communication redundancy. NOAA should integrate NEXRAD data sets into GPM calibration and validation efforts. Satellite Sources Satellite sensors mitigate several weaknesses in rain gauge and ground-based radar data sets; thus, the combination of gauge, radar, and satellite precipitation data provides a powerful tool for multiple applications. For example, space-based sensors sample areas where in situ precipitation observations are absent. In addition, they supplement in situ observations in regions where ground-based sites are sparse (most polar-orbiting passive microwave sensors view a swath that spans the equivalent of three to five NEXRAD radar coverages optimally arranged along a satellite path). Furthermore, satellite observations can delineate areas where no precipitation is falling over vast oceanic regions and merged infrared-microwave products have high temporal refresh. Lastly, geostationary imagers can capture data at a rate that is sufficient to monitor vigorous convective activity (imagers view the entire continental United States every 15 to 30

OCR for page 33
Noaa's Role in Space-Based Global Precipitation Estimation and Application minutes with spatial resolutions of 4 to 8 km at nadir,1 and when operated in “rapid scan” mode, these imagers capture snapshots [over a smaller spatial domain] as frequently as each minute). NOAA’s two major sources of space-based precipitation information are visible and infrared data from geostationary satellites and passive microwave data from polar-orbiting satellites. Visible and Infrared Visible and infrared imagery first became available from satellites in the mid 1960s. The approach to inferring precipitation from these observations was initially centered on indirect (or “proxy”) techniques that relate visible and/or infrared observations of brightness or temperature of cloud tops to large-scale, time-averaged convective rainfall amounts over the global tropics (Arkin, 1979). Thus, although colder cloud tops imply relatively higher cloud tops and the potential for heavier rainfall, the correlation is often poor and varies considerably with time, location, and precipitation regime (e.g., convective or stratiform). However, the wealth of high-quality (1-4 km nadir spatial resolution) geostationary visible-infrared data and the frequent temporal sampling (15-30 minutes over the continental United States) are key ingredients for precipitation monitoring. Although cloud-top temperatures, cloud-top heights, and cloud type are not highly correlated with instantaneous rain rates, the correlation is strongest for warm-season convective systems that frequent the eastern and midwestern continental United States. Operational geostationary visible-infrared digital data sets cover the globe between 60 degrees latitude North and South and are routinely available for near-real-time applications. The geostationary constellation is an international collaboration in which very large data sets are exchanged under the auspices of the World Meteorological Organization. This constellation includes Meteosat-8 (Meteorological Satellite 8; 0° E, European Organisation for the Exploitation of Meteorological Satellites [EUMETSAT]2 ), Meteosat-5 (63° E, EUMETSAT), INSAT (Indian National Satellite; 74° E, India), FY-2 (Fengyun-2; 105° E, China), MTSAT/GMS-6 (Multi-Functional Transport Satellite/Geostationary Meteorological Satellite; 140° E, Japan Meteorological Agency), GOES-West (Geostationary Operational Environmental Satellite-West; 135° W, United States) and GOES-East (75° W, United States). Figure 3.1 shows geostationary sensors and their data availability since 1995. Due to sensor evolution and specific country preferences, no two sensors are identical with respect to available channels or spatial and temporal sampling. All geostationary visible-infrared imagers include a minimum of one visible 1 Directly below. 2 0° East is the nadir angle of the satellite, and European Organisation for the Exploitation of Meteorological Satellites is the operator.

OCR for page 33
Noaa's Role in Space-Based Global Precipitation Estimation and Application are being used in current operational or pre-operational NWP models. The empirical techniques adopt several assumptions about two-dimensional rain rate fields and then adjust the humidity or condensed water profiles based on cloud analyses and latent heat profiles from surface precipitation rates. Variational techniques use observation operators and their adjoints9 to project information from space of the analyzed variables (e.g., temperature, moisture, wind) into that of the observations (e.g., precipitation rate, radiance) and back again in a consistent manner. For precipitation assimilation, the observation operator could be a simplified and (regularized) version of the moist physical parameterizations (and a radiative transfer model in case of radiance assimilation) that relates model state variables to the observations. Thus, a major advantage of variational techniques is (1) that they have the ability to assimilate observations that are not the same as the model variables, (2) that the assimilation is consistent with the model physics, and (3) that spatially and temporally heterogeneously distributed observations are optimally treated. The assimilation of precipitation (and cloud) information is fundamentally more difficult than assimilating temperature, humidity, or wind information. Precipitation is a complex meteorological variable that routinely undergoes dramatic spatial and temporal fluctuations that are not fully understood, much less modeled in near real time. This is particularly problematic for sub-grid-scale processes such as convection. Consequently, satellite-derived precipitation measurements are not yet assimilated into NWP forecast models at many forecast centers. The specific difficulties include (1) limited NWP model ability to accurately forecast quantitative precipitations; (2) inadequate moist physics for clouds, convection, and sub-grid-scale precipitation (retrievals and radiance assimilation are constrained by model microphysics) that includes difficulties relating observed variables to the model variables linked to precipitation; (3) nonnormal observation and background error distributions; (4) non-instantaneous sampling of rapidly evolving rain fields that introduces temporal errors in the data sets; (5) poor knowledge of the statistical properties of clouds; (6) difficulty validating satellite precipitation retrievals; (7) inability to accurately map three-dimensional rain rate structure and fully understand resultant latent heating profiles; and (8) lack of sensitivity of the measurements to drizzle and snowfall. Although the list of hurdles is daunting, progress is feasible through a well-supported, coordinated, multiyear approach spanning several disciplines. 9 For this example, the adjoint is the matrix transpose of the observation operator (tangent linear version of the moist physical parameterization model).

OCR for page 33
Noaa's Role in Space-Based Global Precipitation Estimation and Application Land Surface Land Data Assimilation Systems (LDAS), which use available observations to modify model background fields, can provide a more accurate and unbiased evaluation of initial land-surface states by reducing the accumulating biases in coupled system forecasts of moisture and energy reservoirs.10 Thus, LDAS is similar to atmosphere data assimilation in that it uses available observations to modify a short-range model forecast to provide initialization for a forecast run. Remotely sensed precipitation and soil moisture information is used increasingly in LDAS that may be coupled with atmospheric NWP models (e.g., Rizvi et al., 2002; Drusch et al., 2005). Small-scale spatial and temporal variations in precipitation and available energy, combined with land-surface heterogeneity, cause complex variations in processes related to land-surface hydrology. Characterization of the spatial and temporal variability of the terrestrial water and energy cycles is critical for an improved understanding and modeling of land-atmosphere interaction and the impact of land-surface processes on climate variability. The reservoir and profile of soil moisture and the surface heat balance are the crucial controlling elements for land-surface hydrological processes. Although land-surface layer “wetness” can be inferred from space-based measurements, the total reservoir and profile of soil moisture cannot be directly determined from existing space-based observations. Except for a few specialized local networks, whose soil moisture observations are primarily of value for localized monitoring and development of land-surface models, the total reservoir and profile of soil moisture cannot be determined directly from surface observations either. It is essential to address the observational deficiency of total reservoir and profile of soil moisture to provide information needed for applications such as river stage forecasts, drought monitoring, and crop yield outlooks. Semiempirical land-surface models have been developed to quantify and monitor surface hydrological conditions. These models provide indirect estimates of soil moisture by partitioning precipitation input between surface storage (snow water content), soil moisture recharge, evapotranspiration, and surface and subsurface runoff. Land-atmosphere interactions influence weather and climate variability on a variety of spatial and temporal scales. Because an accurate knowledge of these processes and their variability is important for weather and climate predictions, most forecast centers have incorporated land-surface schemes in their NWP models. Unfortunately, biases develop in model-generated water and energy storage that can continue to grow in the closed, internally cycled, coupled model forecast system. Because these biases can negatively affect forecast accuracy, the NWP 10 See http://ldas.gsfc.nasa.gov.

OCR for page 33
Noaa's Role in Space-Based Global Precipitation Estimation and Application community has been motivated to impose ad hoc corrections to the land-surface states to limit this drift. The U.S. program for the development, application, and improvement of LDAS components is being led by scientists at NASA’s Goddard Space Flight Center and NOAA’s NCEP, in collaboration with researchers at Princeton, the University of Washington, and the NWS Office of Hydrologic Development. This program is focused on the development and application of an LDAS for North America (Mitchell, 2004) and an LDAS for global applications (Rodell et al., 2004). A blended precipitation product is used for the North America LDAS. For the United States, this product is derived by combining 3-hourly precipitation from the NCEP regional model with hourly Doppler radar precipitation and daily rain gauge precipitation. For Canada and Mexico, only the regional model output is used. Several of the current high-resolution satellite-based global precipitation analyses (e.g., high-resolution precipitation products) are being used to force the global LDAS and validate precipitation. NCEP’s Environmental Modeling Center runs the LDAS uncoupled to any atmospheric model and participates in a collaboration on the global LDAS with other agencies. Operational Application of Precipitation Assimilation Techniques Despite the difficulties of assimilating precipitation, a few centers, such as NCEP, Japan Meteorological Agency (JMA), and the European Centre for Medium-range Weather Forecasts (ECMWF) are assimilating precipitation information operationally. In the NCEP regional analysis (i.e., North American Model), precipitation estimates from SSMI, TRMM Microwave Imager, and rain gauges and ground-based radar (over the continental United States) are assimilated using a nudging technique. The analysis has a real-time data cutoff of 45 minutes after the analysis time (e.g., 45 minutes after each 6-hourly model run at 00:00, 06:00, 12:00, and 18:00 Coordinated Universal Time). All data to be assimilated must arrive prior to the data cutoff time. The model temperature, water vapor, and cloud liquid water profiles are adjusted over a 6- to 12-hour window so that the model recomputed rainfall matches the observed (Y. Lin, NOAA NCEP, personal communication, 2006). The NCEP Global Forecast System has a real-time data cutoff time of 2 hours and 45 minutes. The TRMM Microwave Imager and SSMI rainfall estimates over land and ocean are averaged at a 1-degree resolution, and a transformed rain rate is then assimilated variationally, with an assigned observation error that differentiates between land and ocean. The assimilation process changes the temperature, moisture, cloud water mass, and horizontal wind fields. In the Global Forecast System, precipitation assimilation primarily reduces excessive rain rates and, to a lesser extent, increases light rain rates (R. Treadon, NOAA, personal communication, 2006). The impact is greater over oceans than

OCR for page 33
Noaa's Role in Space-Based Global Precipitation Estimation and Application land. Overall, the forecast impact is difficult to quantify with several upgrades combined and tested at once in addition to precipitation, although some relative improvement is shown in the 0- to 24-hour predicted tropical rainfall, forecast wind fields, and tropical cyclone track prediction. NCEP has noted that clear-sky radiance assimilation has a greater impact on forecast precipitation than rain rate assimilation (R. Treadon, NOAA, personal communication, 2006). Given this, plus the ability of the new Joint Center for Satellite Data Assimilation Community Radiative Transfer Model to simulate radiances in cloudy and precipitating fields of view, NCEP plans to move to the direct assimilation of precipitation-affected radiances (R. Treadon, NOAA, personal communication, 2006). Other potential improvements may come through better characterization of the observation and background errors; implementing flow-dependent background error covariances; new analysis systems (GSI and eventually 4D-VAR), and improved model physics and convective scheme (used for assimilation) (Lord, 2004). Lord (2004) also addressed the importance of bias-correcting the observations so that they are consistent with the simplified convective scheme used for assimilation. At ECMWF, precipitation-affected radiance assimilation was recently added to the operational forecast suite (Bauer, 2005). A one-dimensional variational (1D-VAR) retrieval is used to obtain temperature and moisture profiles from TRMM Microwave Imager and SSMI radiances in clouds and precipitation over oceans. From the 1D-VAR retrievals only moisture is subsequently assimilated as total column water vapor in their global, four-dimensional variational analysis (4D-VAR) system. ECMWF plans to move to direct assimilation of precipitation-affecting radiances in its 4D-VAR system in 2007 (Bauer et al., 2006a,b). At JMA, the Radar-AMeDAS (dense network of surface observations including precipitation) composite precipitation data are used in the regional and mesoscale models (Kamiguchi et al., 2005; JMA, 2006). Doppler radar radial wind and precipitable water and rain rate derived from the microwave radiometer on SSMI, TRMM Microwave Imager, and Aqua AMSR-E are used in the mesoscale model. Precipitation information is assimilated using the adjoint of the moist physics that includes both large-scale condensation and convective adjustment (Sato et al., 2004). Even though precipitation assimilation is now operational at several NWP centers, much basic research is still needed to fully exploit the observations. This is discussed in Chapter 4 in the context of NOAA preparations for exploiting GPM data. Monitoring Location and Intensity of Tropical Cyclones and Severe Storms Upper-level clouds commonly prevent satellite data analysts from accurately determining tropical cyclone location and intensity using visible-infrared imag-

OCR for page 33
Noaa's Role in Space-Based Global Precipitation Estimation and Application ery from geostationary and polar-orbiting sensors. Fortunately, some microwave frequencies respond sharply to the frozen hydrometeors and heavy rain characteristics of intense tropical cyclone convective rainbands and eyewall development. Large brightness temperature contrasts permit microwave imagers and sounders with 85-91 GHz channels (i.e., SSMI, SSMIS, TRMM Microwave Imager, AMSR-E, AMSU-B) to supply vital tropical cyclone information (Lee et al., 1999, 2002; Hawkins et al., 2001; Simpson, 2003). Multiple U.S. centers use the passive microwave sensors operationally for tropical cyclone structural details. NOAA’s Tropical Analysis and Forecasting Branch and Satellite Analysis Branch provide the National Hurricane Center with storm location and intensity values for all storms in the Atlantic and Eastern Pacific using microwave products created and distributed by the Naval Research Laboratory and the Fleet Numerical Meteorology and Oceanography Center. The storm-centered microwave products are updated within 1-3 hours of satellite data acquisition and are available worldwide.11 DOD’s Joint Typhoon Warning Center provides multiday forecasts for all storms in the Pacific Ocean, Indian Ocean, and Southern Hemisphere where approximately 80 systems per year typically occur. The Joint Typhoon Warning Center has dedicated satellite analysts who provide the typhoon duty officer with routine storm position and intensity estimates for all active systems. In addition, DOD’s Air Force Weather Agency provides backup resources and creates storm fixes for both the Joint Typhoon Warning Center and the National Hurricane Center using these data sets. The ability to understand storm temporal structure changes via rainband and eyewall configuration trends is crucial to catching storms undergoing rapid intensification, concentric eyewall cycles, and shear and cannot be done with visible-infrared data alone. NCEP plans to implement an extension of the Weather Research and Forecasting Model for hurricane track and intensity forecasting operationally in 2007. The Hurricane-WRF (HWRF) system couples a wave model, an ocean model, a land-surface model, and an atmosphere-ocean boundary-layer model. One of the most significant modeling challenges to improving numerical forecasts of hurricane structure and intensity in high-resolution hurricane models is the initialization of the hurricane vortex. To advance this effort in HWRF, NOAA’s Environmental Modeling Center is developing situation-dependent background error covariances that will be incorporated into a local data assimilation scheme. The immediate goals are to assimilate real-time Doppler radar data from reconnaissance aircraft and coastal WSR-88D radars near land. Future plans call for assimilation of radar reflectivity data and precipitation-affected radiances. 11 Available at http://www.nrlmry.navy.mil/tc_pages/tc_home.html and http://152.80.49.216/tc-bin/tc_home.cgi.

OCR for page 33
Noaa's Role in Space-Based Global Precipitation Estimation and Application POTENTIAL APPLICATIONS OF GPM DATA The availability of GPM-like data from both the TRMM Precipitation Radar and the TRMM Microwave Imager and access to operational and research microwave sensors (SSMI, AMSR-E, WindSat, SSMIS, and AMSU-B) allow realistic predictions of GPM applications now. In addition, they allow testing and refining of methodologies in advance of GPM core satellite launch. Access to near-real-time digital data sets, computer processing power, and automated software has created a wealth of web-based products and databases that feed a diverse and growing user community. Precipitation information serves both near-real-time and non-real-time functions in NOAA operations. Near-real-time applications of satellite precipitation estimates include (1) data assimilation into global and regional models (e.g., hurricane forecasts), and (2) merged or blended global precipitation products (e.g., CMORPH). Non-real-time applications include (1) cross-calibration of GPM’s precipitation radar with other satellite passive microwave precipitation estimates, (2) seasonal models for drought and hydrologic forecasting, and (3) climate modeling. GPM data users are in three principal areas of application: weather prediction and monitoring, hydrology, and climate. This section describes potential applications of GPM data in an operational environment, focusing specifically on NOAA (the committee’s second task). Numerical Weather Prediction Data Assimilation The ultimate goal of NWP is to create useful and accurate forecasts, especially for inclement weather. Areas with clouds and precipitation are often dynamically active, and subsequent NWP forecasts are often sensitive to initial conditions in these areas. Thus, improving their initial conditions should improve the downstream forecasts. Improvement in the initial conditions of the models will result from improved precipitation estimates from GPM and other sensors calibrated by GPM. Moreover, few conventional observations are available for these areas, and satellite data assimilation is essential. Microwave imagers and sounders have improved the ability to monitor near-real-time clouds and rain regionally and globally, but parallel efforts to incorporate these data sets within NWP models have not kept pace because of the complex nature of precipitation variation. Availability of GPM data will help in this regard. The GPM mission has advantages over the TRMM mission for data assimilation and NWP: the improved precipitation rate estimates of the dual-frequency precipitation radar on the GPM core satellite (particularly for light rain) (see Chapter 2) and the increased sampling capacity of the 65-degree inclined orbit of the GPM core

OCR for page 33
Noaa's Role in Space-Based Global Precipitation Estimation and Application satellite.12 Although the swath width of the GPM precipitation radar may limit its utility for assimilation, the relative accuracy of the precipitation radar makes it valuable for validation and calibration of rain rate algorithms for microwave imagers. Even though operational NWP models do well predicting the clouds associated with large-scale organized systems, quantitative precipitation forecasts generally have limited utility beyond 2-3 days. Two contributing factors are that the NWP models more accurately reproduce large-scale motions than cloud microphysics, and the subsequent precipitation forecasts are sensitive to less accurately analyzed or forecast fields, such as vertical motion. The high-precision precipitation radar measurements of horizontal and vertical cloud structure and microphysical elements will provide useful validation data for the development of improved model physical parameterization schemes. Assimilation of the microwave imager precipitation information may help improve the initial conditions required for accurate short-range forecasts of precipitation. High-quality precipitation data are also required for validation of precipitation forecasts. The precipitation radar on GPM will be valuable for this purpose because it provides one of the few sources of precipitation validation data over oceans. The extensive, international ground validation data will also prove useful for validation of NWP precipitation forecasts. To realistically specify real-time rain rates in NWP model analyses, definition of rain rate profiles is critical. Without this key three-dimensional information, precipitation data assimilation must rely on less effective methods. For example, although rain rates can be inferred from microwave imager and sounder data, only radar can actually retrieve vertical profiles. The TRMM precipitation radar is the first satellite sensor to demonstrate that rain rate profiles are feasible from satellite observations (Chapter 2), and the dual-frequency precipitation radar on the GPM core satellite will continue this unique data stream. The radar swath widths of TRMM and the GPM core satellite do not provide the coverage required for nowcasting and/or NWP initialization efforts. Nonetheless, precipitation radar data have been useful for NWP model validation. For example, Benedetti et al. (2005) used precipitation radar data to validate the assimilation of TRMM Microwave Imager and SSMI precipitation-affected radiances into the ECMWF model. The GPM constellation not only has the potential to monitor near-real-time rain events, but with proper planning and support, the data can have an impact on NWP advancements. 12 Compared to TRMM’s orbit between 35 degrees latitude North and South. Details about the GPM precipitation radar frequencies and swath widths are presented in Chapter 1.

OCR for page 33
Noaa's Role in Space-Based Global Precipitation Estimation and Application Monitoring Tropical Cyclones and Severe Storms In conjunction with the sensors on the constellation spacecraft, the microwave imager on the GPM core satellite will have resolution similar to the TRMM Microwave Imager and superior to the SSMIS and microwave sensors flying on the DMSP and NPOESS operational satellites in the 2010-2015 time frame. Thus, the GPM Microwave Imager rainfall data will be crucial in mapping tropical cyclone rainfall, updating R-CLIPER values for future applications, and addressing the global need to monitor potential flooding disasters caused by landfalling tropical cyclones. Flood fatalities are especially likely in developing countries that have large coastal populations, low elevations, and little weather infrastructure to warn about impending heavy rains (Hossain and Katiyar, 2006). GPM near-real-time data sets could be key ingredients to life-and-death evacuation decisions. In addition, GPM will have a direct and measurable positive impact in supporting NOAA’s goal to protect life and property from tropical cyclone damage in the continental United States and countries that rely on U.S.-derived information by improved forecasts of landfall location and intensity. Microwave measurements from GPM’s precipitation radar and microwave imager will give analysts cloud-free views of tropical cyclone structure. Near-real-time information on storm location and structure (highly correlated with intensity) from similar sensors on TRMM has proven operationally beneficial at the National Hurricane Center (NRC, 2004). Hydrometeorological Applications Data from the GPM core and constellation satellites will be valuable for a wide range of applications in hydrometeorology and oceanic meteorology.13 These applications evolve primarily from spaceborne precipitation estimates as well as remotely-sensed soil-moisture and ocean-salinity information. Because of the broad geographic coverage of the GPM mission, GPM data will be particularly valuable in areas of poor ground-based rainfall data for inferring hydrological variability and improving multisensor quantitative precipitation estimates. By underpinning high-quality, high-resolution precipitation products, GPM will support soil-moisture estimates that feed flood forecast models. Such models rely on precipitation estimates and surface radiative balance as the fundamental variables to calculate soil-moisture content in semiempirical land-surface 13 Hydrometeorology, according to the American Meteorological Society Glossary of Meteorology (2nd edition, 2000), is the “study of the atmospheric and terrestrial phases of the hydrologic cycle with emphasis on the interrelationship between them,” while oceanic meteorology relates to the “study of the interaction between the sea and the atmosphere.”

OCR for page 33
Noaa's Role in Space-Based Global Precipitation Estimation and Application model schemes (e.g., LDAS), where precipitation is partitioned between runoff and soil-moisture recharge. Such analyses are particularly valuable for land areas where ground-based precipitation observations are inadequate. Recent examples of successful application of satellite-based precipitation to hydrologic modeling for streamflow forecasting indicate opportunities for GPM applications (see Yilmaz et al., 2005; Moradkhani et al., 2006). The typical spatial resolution of global, satellite-based, high-resolution precipitation product analyses is comparable to that of LDAS components (i.e., 10 to 20 km). The ability to use these products in an operational setting will depend on their availability in the time frame of forecast cycle operations. GPM precipitation data have the potential to contribute to both the further development and the operational application of regional and global LDAS components through improved input and validation of precipitation. GPM data will likely have the greatest impact on global LDAS by improving the quality of the high-resolution precipitation analysis products for the large land areas of the world where surface observations are sparse or not available in a timely manner. Two satellite missions that will potentially overlap with GPM—GOES-R and Aquarius—could enhance the contribution of GPM data to a number of hydrometeorological applications. For example, GOES-R data will be available at 5-minute intervals. In combination with infrared-microwave precipitation estimates with radar-calibrated GPM constellation data, these data will benefit flash flood forecasting. Freshwater input to the ocean surface (by means of precipitation) and ocean salinity are two other potential applications in which GPM data will be useful. The Aquarius mission will measure sea-surface salinity and is a joint project between NASA and the Space Agency of Argentina (Comisión Nacional de Actividades Espaciales). Combining sea-surface salinity measurements14 with inferred precipitation from GPM will be helpful in closing the marine hydrological budget. Furthermore, Aquarius could be a validation tool, albeit an indirect one, for NOAA’s estimates of precipitation over the ocean. While the satellite-based precipitation estimates may have achieved reasonable accuracy to be useful for many purposes, the soil moisture (or more appropriately labeled “surface wetness”) and ocean salinity estimates from spaceborne platforms are still highly experimental at this point (and thus likely not as useful in a quantitative sense) and only available with limited coverage in space and time. 14 The surface salinity depends on air-sea freshwater fluxes from precipitation and evaporation, as well as advection and mixing in the upper ocean.

OCR for page 33
Noaa's Role in Space-Based Global Precipitation Estimation and Application Climate Applications Precipitation Climatology The primary requirements for satellite-related precipitation data for climate applications are high absolute accuracy and long, stable, and consistent time series. GPM would contribute to such applications by extending the time series that began with the launch of TRMM. The limitations of the precipitation climatology developed from TRMM data include (1) a short record (to date); (2) a coarse time-space sampling (i.e., the mission was designed to resolve five-by-five-degree monthly averages); and (3) geographical limitations (35 degrees latitude North and South). Nevertheless, because of the previous meager level of knowledge, TRMM Microwave Imager data have been extremely valuable for describing the large-scale features of the tropical and subtropical precipitation regimes, including individual realizations of the El Niño-Southern Oscillation cycle. In addition, TRMM precipitation radar data have revealed the synoptic climatological features of the precipitation patterns associated with the evolution of tropical cyclones. GPM offers the opportunity to significantly extend the length and geographic coverage of intercalibrated precipitation climatology. Precipitation Analyses GPCP precipitation analyses, along with a new generation of high-resolution precipitation products, exploit passive microwave data from the existing array of polar-orbiting satellites to extend the analysis domain into the high latitudes. As noted earlier, high-resolution precipitation product analyses are primarily used for monitoring and nowcasting weather and short-term climate variability. Some of these analysis schemes exploit TRMM data for intercalibration of passive microwave observations, but the lack of TRMM-type core satellite observations in higher latitudes limits intercalibration outside the tropics and subtropics. GPM will extend these limits from 35 degrees to 65 degrees latitude North and South (the midlatitudes). Efforts are under way at the Climate Prediction Center to extend the short CMORPH time series, which began in December 2002, backward to include the entire period since 1998 when TRMM data became available. It is doubtful whether passive microwave sampling prior to that time is sufficient to generate reliable global precipitation analyses using the CMORPH method (J. Janowiak, NOAA, personal communication, 2006). Further backward extension of the time series would have to rely heavily on infrared-derived precipitation estimates and would almost certainly result in a discontinuity in the time series and poorer-quality analyses. Thus, the length of the high-resolution global time series derived from satellite passive microwave data is likely to be limited to the period since 1998 but would be greatly extended by GPM and, in particular, would be greatly helped in terms of intercalibration if TRMM and GPM overlap.

OCR for page 33
Noaa's Role in Space-Based Global Precipitation Estimation and Application Global Precipitation Climate Data Records The construction of continuing, stable, and accurate global precipitation analyses from observations from a continually changing international constellation of research and operational satellites is a fundamental but daunting challenge for the development of satellite-based global climate data records.15 Fulfillment of this task depends on a number of factors, of which the following two are central: Climate data records depend on the maintenance of an international constellation of satellites that provide the routine time-space sampling needed for high-quality global precipitation analyses. For example, the design sampling interval of the GPM constellation of passive microwave sensors is 3 hours. International coordination to improve the phasing of polar-orbiting satellites could be a factor for improving sampling. Climate data records depend on an ongoing and extensive international program for intercalibration of satellite radiances. Without intercalibration to remove biases, and an understanding and quantification of the error characteristics of the observations, the measurements are of marginal value for climate applications, since drifts in satellite sensors can produce spurious trends in the time series, and jumps can occur in a time series due to systematic biases from different sensor observations. The planned observational and calibration-validation programs of the GPM mission will address many of the current deficiencies that limit the development of satellite-based global precipitation climate data records. SUMMARY The best operational uses of GPM data at NOAA will be weather forecasting, hydrologic applications, climate applications, and global precipitation climate data records. Prior to the launch of the GPM core satellite, NOAA can initiate improvements in current sources of precipitation data and improvements in data products to enhance the operational benefits of the GPM mission. Chapter 4 outlines additional preparation activities at NOAA for optimal use of GPM data by the launch of the GPM core satellite. 15 A climate data record is a data set designed to serve as a climatological basis for diagnosing and studying year-to-year climate variations and decade-to-decade climate change (NRC, 2000). Intercalibration and data continuity are critical components of climate data records. In contrast to environmental data records, which are produced and generally used in real time, the strategy for the production of climate data records involves repeated retrospective reanalysis and refinement, usually based on additional data and information from multiple sources (e.g., improved algorithms).