WMO/CAS/WWW


FIFTH INTERNATIONAL WORKSHOP on TROPICAL CYCLONES


Topic 0.3: Future satellite instruments and opportunities for tropical cyclones

Title: The Increasing Role of Weather Satellites in Tropical Cyclone Analysis and Forecasting

Authors: Christopher S. Velden1 and Jeffrey D. Hawkins2
1 Cooperative Institute for Meteorological Satellite Studies
at the Univ. of Wisconsin
2 Naval Research Laboratory – Monterey, CA

E-mail: chrisv@ssec.wisc.edu and hawkins@nrlmry.navy.mil
Fax: 608 262 5974



0.3.1: Introduction

The invention of weather satellites in the 1960s has greatly reduced the problems associated with tropical cyclone detection. In our current age of technology, a “surprise” hurricane, such as the Galveston event in 1900, is highly unlikely as tropical cyclones can now be detected from space in any ocean basin. Before the 1960s, it was very possible to “miss” a hurricane that was situated or forming away from shipping routes. By the 1970s, high-resolution geosynchronous satellite images from the visible (VIS)1 and infrared (IR) spectra were available at 3- to 6-h intervals over the hurricane-prone regions of the world’s oceans. Numerous experts collaborated to develop subjective cloud pattern recognition techniques and apply them to satellite-detected hurricanes to estimate the maximum winds. Today, the major global tropical cyclone forecast centers rely heavily on meteorological satellite surveillance. In many cases, the satellite is the only available method for estimating motion and intensity.

This report is not an exhaustive treatise on weather satellites or the radiative transfer theory on which satellite remote sensing is based. It is primarily focused on the more recent satellite sensors relevant to the study and forecasting of tropical cyclones. The emergence of new information contributed by space-borne instruments that sense in the microwave end of the spectrum is described. And most importantly, this report provides the IWTC attendees with a summary and outlook on the potential value of anticipated new satellite instrumentation toward improving future tropical cyclone analysis and forecasting.



1 Acronyms are listed alphabetically and defined in the Appendix


0.3.2: Background

0.3.2.1: Applications of Visible-Infrared (VIS-IR) Satellite Imagery and Derived Products for Tropical Cyclone Diagnostics and Forecasts

Satellite instruments receiving energy in the wavelengths of the visible spectrum see much like our own eyes. What these receivers sense derives mainly from reflection and scattering of solar radiation from cloud tops. Hence, the VIS imagery is not available during the nighttime hours. On the other hand, radiation sensed in the IR wavelengths comes from emission by the clouds, so that the energy emitted by the cloud tops can be received by a radiometer in space around the clock. The cloud-top temperatures (and altitudes) can be estimated if the IR sensors are properly calibrated and an atmospheric temperature profile is available nearby. It should be pointed out that because the IR sensors are “blind” to radiation below opaque clouds, the imagery is often confined to a depiction of the storm canopy and the “exhaust” clouds associated with the upper-level outflow of the hurricane. In conditions where the storm is characterized by a central dense overcast (no apparent eye in the upper-level cloud structure), it is often not an easy task to infer the exact center location and particularly the maximum surface wind speed.

Using this data, NOAA scientist Vern Dvorak developed his subjective image pattern recognition technique, which was first published in 1975 [Dvorak, 1975]. This method, although since refined by researchers such as NOAA scientist Ray Zehr and operational satellite meteorologists in practice, is still in use today as the standard for estimating tropical cyclone intensity around the globe.

The Dvorak technique is based on principles of cloud pattern recognition. The amount of organization is directly related to tropical cyclone intensity. Trained satellite analysts determine the cloud pattern type and relate it to storm development stages. As satellite IR sensors became more sophisticated in the 1970s, Dvorak expanded his techniques to include special enhancements. The Dvorak enhancement curve was developed to allow satellite specialists to focus on the convective vigor in the eyewall. A statistical relationship was developed between the IR-depicted eye (warm) temperature and the surrounding eyewall cloud (cold) temperature. From this relationship, intensity could be inferred. This enhanced IR (EIR) technique was added to the Dvorak scheme. Further details on this landmark study in hurricane meteorology can be found in Dvorak [1984].

Recently, university scientists collaborated to automate the Dvorak technique. This computer-based algorithm was designed to minimize human judgment in cloud pattern typing. While experienced satellite analysts are still required, the Objective Dvorak Technique (ODT) has been shown to be competitive [Velden et al., 1998b], and this algorithm has been transitioned into several regional forecast office operations where it is part of the suite of satellite-based guidance tools. An advanced version of the ODT (AODT) is being developed for experimentation in the near future.

Another satellite tool was introduced in the late 1970s and 1980s. The “water vapor channel” was added to IR sensors on the geosynchronous weather satellites deployed by European nations and the United States. This channel focused on longwave radiation absorbed and re-emitted by water vapor in the 6-7 micron frequency range. In contrast to the IR imagery, it was no longer necessary for clouds to be present as this channel could depict moisture structures in the atmosphere as never seen before. Animation of this imagery revealed circulations and synoptic features in the tropical cyclone environment that would affect the motion and track forecasts. Today, water vapor imagery is a prominent tool for use in analyzing the hurricane environment.

Given the scarcity of conventional observations over the oceanic regions, the satellite-derived winds from these animated image sequences have been embraced by tropical cyclone forecast centers. Despite the limitation of imprecise height assignments, the data are used to help determine the environmental flow field that accounts for much of the tropical cyclone steering and forecast motion. Today, automated techniques are in place that track features in multispectral full disk imagery from multiple geosynchronous platforms to provide global wind sets around the clock. These data sets have proven useful in real-time analysis [Velden et al., 1998a] and in reducing tropical cyclone track forecast error in numerical models [Goerss et al., 1998; Soden et al., 2001]. Special rapid-scan schedules from the GOES weather satellite are now standard and can sample at a 1-5 minute frequency over targeted tropical cyclones, and thus provide very detailed wind fields (Berger and Velden, 2001). GMS and Meteosat satellites can also scan in rapid-scan mode.

Another use of IR information is the determination of sea-surface temperatures (SST). Hurricanes depend on the ocean heat content and moisture fluxes to drive the convection necessary to sustain their low pressures, and it is well recognized that ocean surface temperatures greater than 26_C are usually necessary to generate hurricane vortices. Satellite IR radiometers are able to sense the ocean skin temperatures in the absence of clouds. This information is particularly important as strengthening or weakening could depend on the SST gradients in the path of a hurricane. Features such as the warm Gulf of Mexico loop current, the warm Gulf Stream and associated eddies, and cool-water “wakes” left over from the heat loss and mixing associated with previous storms may factor into the intensity forecasts. Recently, studies have shown (e.g., Shay et al. 2000; Fu and Casenave, 2001) that satellite altimeter data can provide additional information on the ocean heat content (energy available in the upper layer of the ocean).

A major concern associated with landfalling tropical cyclones is the rainfall and flooding potential. IR imagery is used as a guidance tool to infer rainfall rates in hurricanes. Studies have been directed towards calibrating satellite IR-based rainfall estimation techniques to radar estimates at landfall. While these IR methods can provide a generally good (within a factor of two) bulk estimate of rainfall in the absence of other data, precise observations of heavy rain bands and localized accumulations are difficult from the IR alone. Improved multispectral satellite techniques are being developed that utilize the advantages of rapid temporal sampling from geosynchronous IR imagery and the cloud-penetrating properties in the microwave spectrum from polar-orbiting instruments (Karyampudi et al. 1999; Turk et al. 1999).

IR and VIS satellite observations have become a mainstay of tropical cyclone detection and analysis. The limiting factor of passive remote sensing in these spectra is the inability to sense through clouds. The next section will address this limitation with the introduction of new satellite sensors that can penetrate clouds and provide a unique view of tropical cyclones.


0.3.2.2: The Emergence of Observations in the Microwave

Satellite-based passive microwave imagery has been shown to enhance our ability to monitor key hurricane characteristics [Hawkins et al., 2001]. Passive microwave data ameliorates the prime limitation inherent with VIS/IR data by being able to sense through non-raining clouds and mapping organization not otherwise detectable. The longer wavelengths associated with microwave radiation are not sensitive to thin ice clouds and can therefore penetrate the storm’s cirrus canopy to reveal important structure underneath. Unfortunately, the longer wavelengths require larger receivers that currently limit the instrumentation to low-altitude, earth-orbiting satellites. Therefore, very frequent sampling (which is an advantage of geosynchronous satellites) and image animation are not currently possible.

The concept of estimating tropical cyclone structure, intensity and intensity change using polar-orbiting weather satellite passive microwave observations spans over two decades of active research. Recently, progress has been made in part due to advanced microwave sounding instrumentation. An excellent example of this new and improved capability is the NOAA-15 Advanced Microwave Sounding Unit (AMSU) described by Kidder et al., [2000]. Temperature profiles derived from AMSU measurements can be used to construct vertical cross-sections of the tropical cyclone core region warm anomaly (Figure 1).



Figure 1. NOAA-15 AMSU vertical temperature anomaly (oC) for Hurricane Floyd at 1238 UTC September 14, 1999 (right). Floyd’s warm core is clearly evident with temperature anomalies in excess of 18oC near 250 hPa. Strong cooling (>-16 oC) below the convective eyewall region is fictitious and due to the lack of explicit correction for scattering in the sounding retrieval package. AMSU 55.5 GHz, 54.94 GHz, 54.4 GHz and 53.6 GHz limb corrected brightness temperatures (Tb, top to bottom) with corresponding peak radiance sampling levels (horizontal lines) are shown with coincident color-enhanced VIS image inset (left). [UWISC/CIMSS]



The ability of the AMSU to detect changes in a tropical cyclone warm core structure is vitally important as it yields information on storm intensity and intensity change. Recent research [Brueske and Velden, 2002] shows promise in further quantifying the relationship between AMSU-observed upper-tropospheric warming and hurricane intensity. In this study, explicit correction of scan-geometry related warm core sub-sampling is modeled. More precise intensity estimates are then possible. Starting in 2001, experimental AMSU-based tropical cyclone intensity estimates were made available in near real-time for evaluation by forecasters at the NHC and the Joint Typhoon Warning Center (JTWC). It is expected that these techniques will be improved and transitioned into operational forecast centers in the near future.

Another unique capability of the AMSU is its ability to provide important information on the structure and distribution of tropical cyclone gale-force winds using thermodynamic and dynamic constraints. This method is described in Demuth et al. [2000]. A statistical algorithm is used to determine the temperature as a function of pressure from the AMSU radiances. The hydrostatic equation is then integrated downward from 50 hPa to the surface using an upper-level boundary condition provided from an NWP analysis. Prior to the hydrostatic integration, an empirical correction is applied to the temperature profiles to adjust for attenuation of the AMSU radiances by liquid water and scattering by ice. Once the height field is determined as a function of pressure, the wind field is calculated assuming gradient balance. The height field is azimuthally averaged relative to the storm center for this purpose. A statistical algorithm has been developed to estimate the maximum winds and the radii of 34, 50, and 64 kt winds from properties of the AMSU analyses. The inputs are AMSU analysis properties such as surface pressure drop from 600 km to the storm center, the maximum retrieved gradient wind, and the average cloud liquid water near the storm center. The output is an estimate of the observed maximum wind and wind radii. This promising algorithm is currently under development and is being evaluated by NHC.

Another important satellite tool became available to forecasters in the late 1980s. The U.S. Department of Defense (DoD) was the first to fly a series of passive multifrequency microwave imager on an operational meteorological satellite in 1987, and this program of satellites continues. The Special Sensor Microwave Imager (SSMI) channels, unlike the NOAA sounder channel data described above, penetrate to the Earth’s surface unless the signal is attenuated by tropospheric hydrometeors. One of the SSMI channels senses at a frequency of 85 GHz, with a spatial resolution of 13-15 km. This channel is able to penetrate non-raining clouds due to the small size of ice crystals aloft. However, larger, frozen hydrometeors (e.g., hail, graupel) and rain drops associated with vigorous convection dramatically scatter radiation at this frequency. Thus, the sensor can easily detect intense rain associated with hurricane rainbands and the eyewall due to the lower “brightness temperatures” created by the intense scattering. A time series of 85 GHz measurements can reveal a storm’s internal convective structure and evolution by mapping the organization and vigor of the convection around the storm center. It is important to note that these data have been enthusiastically used by virtually all of the World Meteorological Organization (WMO) tropical cyclone warning centers over the globe (via the Naval Research Laboratory TC web site: http://www.nrlmry.navy.mil/tc-bin/tc_home

Building on the success of the DoD program, NASA launched a special satellite aimed at measuring meteorological quantities over the tropics using passive and active microwave sensors. The Tropical Rainfall Measuring Mission (TRMM) satellite completed 3 years of successful data taking in 2000. With the change in altitude from 350 km to 400 km made in August 2001, it should have enough fuel to continue measurements until about 2005. TRMM is a low earth-orbiting satellite with a unique orbit that oscillates about the Equator between roughly 35N and 35S. A joint mission between the U.S. and Japan, TRMM’s primary stated goals were: 1) improve the understanding of crucial links in climate variability due to the hydrological cycle, 2) improve the large-scale numerical models of weather and climate through assimilation of TRMM data, and 3) advance our understanding of cloud ensembles and their impacts on larger scale circulations. Shortly after launch, scientists recognized that TRMM would also provide essential new information on tropical cyclones in all stages of development. Kummerow et al. [2000] and Lee et al. [2002] provide more detail on the TRMM instruments, algorithms and a wide range of early results.

The TRMM microwave imager (TMI) is a multi-channel, dual-polarized, conically scanning passive microwave instrument similar to SSM/I. The TRMM TMI complements the SSM/I data by filling in temporal gaps and providing superior spatial resolution. The TMI 85 GHz ground footprints are 5-7 km, which is 2-3 times better than the SSM/I, and the higher resolution TMI 37 GHz channel can penetrate deeper into the tropical cyclone to reveal additional details about the storm organization. The purpose of the VIS/IR instrument (VIRS) is to enable TRMM to be a “flying rain gauge.” The TRMM satellite radar and radiometer combination is intended to obtain high quality vertical profiles of precipitation as well as surface rainfall estimates. TRMM's physically based rainrate observations from the combined radar and passive microwave instruments allow the calibration of empirical rain estimates from the IR sensors. As a result, uncertainty in tropical rainfall has been greatly reduced from earlier space-based estimates. Currently, these techniques are being applied to better estimate tropical cyclone rainfall.
Along with the increased understanding of the global water cycle, the value of TRMM to the forecast and research communities is evident by its wide usage in operational tracking and forecasting of tropical systems, along with the increased understanding of the global water cycle. TRMM data are being provided in near real time to many tropical cyclone forecast centers via the NRL web page. The high-resolution imagery is used to help locate tropical cyclone centers and to assess convective organization trends.

The ability of passive microwave data to sense through high-level cirrus is especially relevant to detect lower-level features that define the circulation center, which can lead to more accurate storm center fixes. Figure 2 illustrates how an exposed low-level circulation center (LLCC) is readily apparent in the SSM/I product, while the cirrus shield in the enhanced IR image obscures this feature. Use of the SSMI and TMI can help reduce large TC center location errors.




Figure 2. Comparison of color-enhanced SSM/I product (left) with the standard Dvorak IR enhancement (right) for tropical cyclone 05S on January 16, 2001 at 0149 UTC. Both images have 2_ latitude/longitude markings and reveal the ability of the SSM/I to detect sheared systems with exposed low-level circulation centers (LLCC) not viewable in IR. [Naval Research Laboratory – Monterey, CA]



Passive microwave data are also quite valuable when trying to understand the organization of systems undergoing rapid intensity changes. Intense tropical cyclones are often characterized by double (or concentric) eyewalls. Double eyewalls are rarely seen in VIS/IR imagery since upper-level clouds almost always obscure these structures. However, the passive microwave sensitivity to large ice particles and rain scattering in the inner and outer eyewalls typically allows this important structural signature to be observed. A good example of this phenomenon is illustrated in Figure 3, which shows coincident TMI and geostationary satellite IR data during Hurricane Gert as it underwent a double eyewall formation. The IR data caught the formation of the eye due to the dissipation of high clouds from subsiding air in the eye. The outer eyewall is clearly shown in the TMI image, but is obscured under the cirrus canopy in the IR image. Trained forecasters can observe trends in these data (strong and constricting inner eyewall generally means intensification; collapsing inner eyewall in favor of the outer eyewall generally implies decreasing intensity) and use these trends as a proxy for the evolution of eyewall dynamics. This information can serve as important guidance in their intensity forecasts.




Figure 3. TRMM 85 GHz color product detecting inner and outer eyewall for Hurricane Gert (left), while Dvorak IR enhancement shows little evidence of outer eye (right) due to obscuration by upper-level clouds. [Naval Research Laboratory, Monterey, CA]



Since the microwave-sensed signal is a result of scattering by hydrometeors, it becomes an immediate candidate for tropical cyclone rainrate estimation techniques and does offer some advantages over IR-based methods. Combined microwave and IR techniques are being developed and calibrated with radar estimates in cases of landfalling tropical cyclones. Experimental versions are showing promise, and should lead to improved operational rainfall estimates in the near future. The importance of accurately estimating tropical cyclone rainrates cannot be understated, as inland flooding is a leading cause of human fatalities.
Satellite-borne scatterometer data can be an important analysis tool for tropical cyclone forecasters (Hawkins and Black, 1983). A review of scatterometry and science applications is provided by Liu [2002] and Edson and Hawkins [2000]. A scatterometer sends microwave pulses to the Earth's surface and measures the backscattered power from the surface roughness elements. Over the ocean, the roughness elements producing the backscatter are largely due to small waves on the surface, which are believed to be in equilibrium with the local wind stress. The backscatter power depends not only on the magnitude of the wind stress, but also on the wind direction relative to the direction of the radar beam. Wind vector retrievals from backscatter measurements are based on empirical relations. The relation between backscatter and ocean surface winds under the strong wind and rainy conditions of a hurricane is not well established because of the lack of validation data and sudden failure of the early NASA scatterometer (NSCAT). The ability to estimate surface wind speed and direction under both clear and cloudy conditions day and night is what makes the scatterometer unique.


The success of earlier scatterometers and sudden failure of NSCAT prompted the rapid deployment of a new instrument by NASA in 1999. Called QuikSCAT, the conical scan of this sensor’s pencil-beam antennas provides a continuous 1800 km swath and unprecedented global ocean coverage. Wind fields from QuikSCAT are available in near-real time to most tropical cyclone forecast centers. The standard wind product has 25 km spatial resolution, but special products with 12.5-km resolution have been produced to monitor tropical cyclones. Figure 4 illustrates an example of QuikSCAT wind data during Hurricane Floyd.

In examining thousands of collocated QuikSCAT and ocean buoy measurements, researchers found almost no mean difference (bias) in wind speed and only 6° variance in wind direction (assuming you pick the best wind direction out of the potentially four available). However, a debate is ongoing in the tropical cyclone and satellite research communities over the accuracy of scatterometer wind retrievals under the extreme wind and rain conditions associated with hurricanes. Because of the lack of in situ wind measurements, the accuracy of the wind retrievals under high wind conditions (over 15 m s-1) and in heavy precipitation is more uncertain. There are clear indications that heavy rain affects the wind retrieval, but there is currently no proven threshold of rain intensity at which the wind retrieval becomes invalid using existing algorithms. This is an area of intense research (Katsaros et al. 2001; Edson et al. 2002). Currently, wind vectors under suspected areas of rain contamination are flagged, and the forecaster is advised to use these data with caution. Nonetheless, the satellite scatterometer surface wind vectors have become an integral part of the tropical cyclone analysis toolbox being utilized by operational forecast centers.




Figure 4. Example of QuikSCAT surface wind data overlain on TRMM rainfall signatures during Hurricane Floyd in 1999. The length of the wind vectors is proportional to the wind speed. [Taken from Liu et al., 2000]



0.3.2.3: Summary

The last half of the 20th century bore witness to ‘the burgeoning role of weather satellites’ in monitoring tropical cyclone activity. From a global view, satellites have become the predominant tropical cyclone monitoring platform. Analysts have developed methods to interpret the imagery to infer intensity and short-term motion. Newly available passive microwave imagery has provided unique insights into the organizational structure of hurricanes. The ability to observe beneath the tropical cyclone cloud shield has enabled researchers and forecasters to increase their understanding of fickle storm tendencies.
Yet the fact remains that the satellites yield only remotely sensed data. They do not ‘feel’ the tropical cyclone elements, but ‘see’ them from afar. Despite the recent advances in instrumentation, satellite data must be augmented by conventional in-situ observations to achieve an optimal monitoring scenario. Reconnaissance aircraft still provide the highly reliable inner-core data in Atlantic basin hurricanes that are needed for precision. An unfortunate consequence of improving satellite technology was the cost-cutting discontinuance of aircraft reconnaissance missions into western North Pacific typhoons in 1987. Not only was this a loss of valuable real-time data for prediction purposes, it also cost researchers an important dataset for validation studies. Over time, it is hoped and believed that advances in remote sensing can reduce the risky and costly aircraft flights into tropical cyclones. But until weather satellite sensors can achieve the accuracy of in situ measurements, it is vitally important that aircraft-based tropical cyclone reconnaissance missions (manned airplanes and/or unmanned drones) continue.

The success of the Tropical Rainfall Measuring Mission (TRMM) has led to the formulation and high prioritization of a Global Precipitation Mission at both NASA and its Japanese counterpart NASDA. A tentative launch date of 2007/8 is being planned. The mission is envisioned to consist of a core satellite carrying a dual-wavelength radar and a TMI-like radiometer, and is expected to achieve an incremental improvement in precipitation measuring accuracy from space-borne observations. In cooperation with the international community, NASA plans a constellation of up to eight passive microwave instrument-bearing satellites flying at about 600 km in sun-synchronous orbits. With these programs, the sparse sampling of TRMM will be overcome, and more frequent observations of tropical cyclones will be possible. The obvious synergy between the core satellite, which is designed to improve our physical understanding of rainfall systems, and the constellation satellites designed to improve sampling implies that the Global Precipitation Mission will deliver high quality measurements with the high temporal frequency needed by many applications requiring accurate rainrate data.


NOAA and the DoD are merging the NOAA and DMSP series of satellites into one program, the National Polar-orbiting Operational Environmental Satellite System (NPOESS). This new series will build upon the successes of the Advanced Very High Resolution Radiometer (AVHRR), AMSU, Operational Linescan System (OLS), SSMI and Special Sensor Microwave Temperature (SSM/T1 and SSM/T2) measurements. These satellites will contain advanced instrumentation with improved spatial resolution. With an improved global water cycle budget, the complex interactions between Earth’s biosphere, atmosphere, oceans, and cryosphere will be better understood and simulated in prediction models. Most directly important to the public, tropical cyclone forecasts should be improved through earlier warnings.


Based on preliminary success and scientific results of GPS/MET radio occultation studies applied to measuring atmospheric variables, a team of international partners is developing COSMIC (Constellation Observing System for Meteorology, Ionosphere, and Climate), which is a follow-on project for weather and climate research, climate monitoring, space weather, and geodetic science (Rocken et al. 1997; TAO 2000). The COSMIC program plans to launch a small constellation of six low-Earth orbiting satellites in 2005. Each COSMIC satellite will retrieve about 500 daily profiles of atmospheric thermodynamic properties from the tracked GPS radio signals as they are occulted behind the Earth limb. The horizontal spacing of COSMIC data on a daily basis varies from approximately 400 km near the poles to 500 km in the tropics. Thus, COSMIC data by themselves will not resolve the detailed structure of tropical cyclones with typical scales of several hundred kilometers. However, they will contribute significantly to the definition of the pressure, temperature, and water vapor, and through data assimilation, the wind field in the large-scale environment surrounding the storm. These improved analyses and resulting numerical model forecasts should provide significant benefits to tropical cyclone research and prediction.


0.3.3: Outlook: Applications of Current and Future Satellite Systems to Tropical Cyclone Analysis and Forecasting

Satellite observations of tropical cyclones can be used for “diagnostic” and “prognostic” purposes. For a diagnostic application, satellite data are used to determine a specific property of a tropical cyclone such as the position, intensity, etc. The Dvorak classification technique is an example of a diagnostic algorithm for storm intensity. For a prognostic application, satellite data are assimilated to more accurately determine the initial condition for a statistical or numerical TC prediction model. Satellite-derived winds are a good example here. We propose that current and planned satellite-derived data will be useful toward both of these types of tropical cyclone applications.

The specific types of future satellite observations with applications to tropical cyclones discussed in the previous sections can be divided into six basic categories: 1) Visible/Infrared Imagers, 2) Infrared Sounders, 3) Microwave Imagers, 4) Microwave Sounders, 5) Scatterometers, and 6) Radar Altimeters. As discussed above, the first two types of measurements are available from geostationary and polar-orbiting satellites. The last four are currently only available from polar orbiting satellites (and this will be true for the foreseeable future). Listed below are some of the satellites/instruments that will provide these measurements (see the Acronym list in the Appendix). We now discuss how these systems can be used for tropical cyclone analysis and forecasting. Although the TC applications are listed separately for each group of measurements, the full value of the satellite observing system will be realized only when the best aspects of all of systems are combined.


1) Visible/Infrared Imagers

Geostationary:
GOES (USA) (current)
GOES N-series (2004-2011)
GOES R-series/Advanced Baseline Imager (ABI) (2012-?)
Meteosat Second Generation/Advanced Imager (Europe) (2002-?)
METSAT (India) (current)
MTSAT (Japan) (2003)
FY-2C (China) (2003)
GOMS-2 (Russia) (2003?)

Polar:
NOAA AVHRR (current-2010)
DMSP/OLS (current-2015)
Terra, Aqua/MODIS (current-2007)
SeaWiFS ( current - ?)
OceanSat (current - ?)
METOP/AVHRR (2005-?)
NPP/VIIRS (2005-2011)
NPOESS/VIIRS (2009-?)

The imagers on the geostationary (GEO) and polar-orbiting (low earth-orbiting (LEO)) satellites provide the data for the position and intensity diagnostic algorithms that complement the aircraft data in the Atlantic basin. In other tropical cyclone basins, position and intensity estimates from satellite imagery are the primary sources of these parameters. The visible and IR Dvorak methods have been used for about two decades, and research is currently underway to refine objective versions of these algorithms (i.e., the Objective Dvorak Technique (ODT) currently being tested at NHC). As additional imager channels with higher resolution become available, new research will be required to improve the position and intensity estimation methods, and develop new methods that are not necessarily constrained by the current Dvorak rules.

Satellite IR imagery is used for tropical cyclone precipitation estimation, which is of particular importance in landfall situations. The geostationary instruments have the advantage of high time resolution, while the polar-orbiting instruments tend to have higher spatial resolution. The GOES hydro-estimator technique developed by NESDIS makes use of the IR window channel, and a newer multi-spectral method uses all five of the current GOES imager channels. The advanced imagers of the future will have many new channels. In the longer ranges, the GOES R-series satellites will provide such opportunities. In the near-term, research in improved rainfall estimation methods could also be performed using new imager channels on polar satellites (MODIS, NPP, etc). These data will provide new opportunities for improving IR-based tropical cyclone rainfall estimates. The bigger payoff may likely be the increased number of passive microwave sensors and their combination with IR sensors.

The primary energy source for tropical cyclones is oceanic heat content. Changes in the upper-oceanic thermal structure can affect the intensity trends of tropical cyclones. Satellite-based Sea Surface Temperature (SST) analyses make extensive use of the AVHRR and GOES IR imager channels, and more recently TMI channels. These analyses have the potential for improvement by taking advantage of new channels which are becoming available to better correct for aerosols (for example, from moisture or African dust). Combined SST analyses making the best use of IR (GEO + LEO) and passive microwave measurements will likely lead to improved all-weather products.

The IR imager channels are also used extensively for determining cloud and water vapor motion winds. It has been shown these data can have a notable positive impact on numerical tropical cyclone track forecasts. These data are also used extensively in a diagnostic mode by forecasters. Recent research using rapid-scan imagery to derive the wind fields in tropical cyclones shows promise. From a diagnostic standpoint, these satellite winds are very useful for improving the representation of the storm environment to reduce track forecast errors, as well as for diagnosing the wind structure at the top of the storm (intensity applications). In addition, high-resolution low-level cloud drift winds from the GOES visible and 3.9um channels have been experimentally incorporated into the NOAA Hurricane Research Division (HRD) H-WIND surface analyses, and often provide crucial information on the extent of outer TC radii wind speeds. This is of particular importance to the landfall scenario as these radii help emergency managers determine the timing of onset and extent of possible storm wind effects. Further research is needed to refine and optimize the assimilation of these winds into H-WIND. In the upcoming years, these wind estimates may also be improved by taking advantage of new IR imager and sounder channels. However, assimilating these wind data into tropical cyclone numerical weather prediction models is a current challenge and in need of immediate research initiatives.

Information derived from the IR imager channels (current and future) has the potential to improve tropical cyclone numerical model forecasts by assimilation of the radiances. Cloudy radiances can be assimilated directly into the model and together with other variables can be used to improve the model precipitation fields. Clear-sky radiances can be assimilated in the tropical cyclone periphery to better establish the environmental conditions. Initial conditions for the model forecasts can also be improved by directly assimilating the satellite precipitation estimates through a procedure called physical initialization, which adjusts the model moisture/precipitation fields to agree with the observations.


2) Infrared Sounders

Geostationary:
GOES -- 18 Channel Sounder (current)
GOES N-series -- 18 Channel Sounder (2004-2011)
IOMI/GIFTS -- hyperspectral (~2006)
GOES R-series/HES – hyperspectral (2012-?)

Polar:
NOAA 15-17/HIRS (current-2010)
AQUA/AIRS -- hyperspectral (2002-2006)
METOP/HIRS (2003-?)
NPP/CRIS (2005-2011)
NPOESS/CRIS (2009-?)

The primary tropical cyclone application of the IR sounder data is through assimilating the radiance information into numerical models. Beginning with AIRS, and continuing with GIFTS and HES, the IR sounders will have a dramatic (two orders of magnitude) increase in the number of channels. This will lead to large improvements in the vertical resolution of the derived temperature and moisture profiles in clear areas and above cloud top level. These measurements would improve the representation of the initial temperature and moisture fields in the storm environment.


The horizontal and vertical resolution of the GIFTS and HES instruments might be high enough so that near-continuous eye soundings could be obtained for storms with well-defined eyes. Diagnostic intensity algorithms could be developed from this data. To better describe the tropical cyclone environment, the high-resolution sounders will have immediate application via improved height assignments for cloud and water vapor motion winds. Research is needed on developing a method for tracking moisture or temperature structures derived from clear-sky sounding fields, where the altitude of the features is already resolved by the retrieval.


3) Microwave Imagers

Polar-only:
DMSP/SSMI (current)
TRMM/TMI (current)
AQUA/AMSR (2002-2006)
ADEOS-II AMSR (2002-?)
WindSat (2002-2005/7)
DMSP/SSMIS (2003-2012?)
Advanced AMSR (GCOM B1) (2007-?)
NPOESS CMIS (2008-202?)
GPM (2008-?)

The primary advantage of microwave observations is that they can provide information below cloud tops. Tropical cyclone applications of microwave imager data include qualitative methods for determining storm position, structure (eyewall, rainbands below the cirrus shield, etc), and for quantitative surface wind speed (outside raining areas), rain rate estimation, total precipitable water, and cloud liquid water. These data have the advantage over IR data for rainfall estimation because the measurements are more physically related to the precipitation. Fruitful areas for tropical cyclone research are to combine the high temporal and spatial resolution precipitation estimates from IR imagery on the geostationary satellites with the more physically based, but lower temporal and spatial resolution of the microwave precipitation algorithms (Turk et al., 1999).


The quantitative use of the microwave imager data for diagnosing tropical cyclone intensity has not been fully exploited. In particular, there is TRMM in recent years, and also a goldmine of SSMI data dating back to 1987. The “radar-like” signal presented by the SSMI microwave imagery differs from the IR in that only active (precipitating) clouds are delineated. This eliminates much of the tropical cyclone storm canopy cloud that can often hide important structure from the IR techniques. Storm organization and contracting eyewall cycles can be much more easily recognized in these data, which can lead to improved intensity diagnoses. Research in this area is highly recommended, especially if fused with existing/proven IR-based objective methods (Bankert and Tag, 2002).


4) Microwave Sounders

Polar only:
NOAA-15-17/AMSU (current-2010)
HSB – Aqua (2002-?)
DMSP/SSMIS (2003-?)
METOP/AMSU/MHS (2005-?)
NPP/ATMS (2005-2011)
NPOESS/ATMS (2009-?)

Similar to the microwave imager data, the microwave sounder data can provide valuable information below cloud-top level. Tropospheric thermal measurements can be obtained in non-raining cloudy regions (although at reduced vertical resolution). One of the largest reduction in errors in global forecast models occurred when NWP centers began to assimilate the radiances from the AMSU instrument. In tropical cyclone applications, the microwave sounder data have proven to be useful for diagnosing storm intensity and size. Current algorithms are limited by the footprint size (50 km near nadir) of the AMSU instrument. However, these algorithms are showing great promise as an alternative to the IR-based methods. Further research (near term) is needed to refine these techniques, and develop integrated algorithms that exploit the microwave sounder with the imager (i.e., SSMIS) and IR. These algorithms can be further advanced as the microwave sounding technology improves (i.e., with the introduction of ATMS).


5) Scatterometers

Polar-only:

QuikSCAT (current-2003)

ADEOS II SeaWinds (2002-?)
METOP/ASCAT (2003-?)
SeaWinds (AlphaSCAT) (2007-?)
NPOESS/ASCAT (2009-?)

The primary application of the scatterometer is for ocean surface wind vectors. These data are already proving to be very useful for determining the outer wind structure of tropical cyclones, identifying closed circulations of developing systems and providing lower limits for maximum sustained winds. Efforts are also underway to assimilate these data into numerical models (ECMWF and NCEP currently assimilate operationally and FNMOC will by 2003). The current QuikSCAT system is limited by the fairly large footprint size and attenuation in heavy rain areas, and can only reliably estimate surface winds up to 40 kt. These limitations should become less of a problem as the scatterometers continue to improved by moving away from Ku-band that is so susceptible to rain. The alternative is to have collocated passive microwave to allow rain correction. Further research on the development of specific tropical cyclone surface wind algorithms is encouraged.


6) Radar Altimeters

Polar-only:
ERS-2 (current)
Topex-Poseiden (current)
Jason-1 (2002-?)
Envisat (2002-06)
Icesat (2002-05)
Cryosat (2004-07)
ISS Abyss (2006-12)
Wittex (2006-10)
Geosat Follow-On (current)
Jason-2 (2007-11)

Many studies have shown that ocean sub-surface thermal structure plays an important role in tropical cyclone intensification. In areas such as the western Caribbean and near warm-core rings in the Gulf of Mexico, the warm water extends to much larger depths than in other areas, so there is much more heat energy available to the tropical cyclone. The sub-surface ocean structure can be inferred from satellite altimetry data. Research to better use this information in statistical forecast algorithms and in coupled ocean-atmosphere models has the potential to improve tropical cyclone intensity forecasts.


The altimeter observations would be more useful with improved temporal and areal sampling. However, this is difficult due to the necessity for repeat tracks required for many applications. Proposed scanning or multi-beam altimeters potentially can dramatically increase spatial sampling and fill in data voids. International cooperation is required and is moving forward. Fortunately, time scales for many oceanic phenomena are much slower than their tropospheric counterparts, so that even limited temporal observations still provide useful information.



0.3.4: Summary

The future appears bright for our space-based observing system. Advanced, multispectral (visible, IR, and passive microwave) imagers, sounders (infrared and microwave) and scatterometers are planned for launch in the near future. Hyperspectral measurements from newly developed interferometers are expected to be flown experimentally by 2006. The information content will vastly exceed that of the current measuring devices. Instead of a few dozen viewing channels, these instruments will have more than a thousand channels over a wide spectral range. The satellite data downloads are expected to exceed several terabytes per day. Fortunately, communications and computing capacity are increasing at a rate that hopefully can accommodate this data explosion. Emerging new technologies (including the use of rapidly-developing visualization tools) will be employed. It is important that the evolving space-based observational system keep one step ahead of the demands being placed by the user community and advances in numerical weather prediction. While it will become an enormous task and challenge to assimilate this wealth of data into meaningful parameters, the outlook is bright for unlocking the still-unresolved mysteries toward improving our understanding and prediction of tropical cyclones.



All of the above-mentioned satellite-based instruments provide unique information by themselves. However, we firmly believe that the sum of the parts holds even greater potential in tropical cyclone applications. We therefore strongly encourage applied research that is focused on fusing these varying data types into specific tropical cyclone algorithms that can cross-calibrate and take advantage of each instrument’s strengths.



Acknowledgments:
We gratefully acknowledge the support of our research sponsors, the Office of Naval Research, Program Element (PE-060243N) and the Oceanographer of the Navy through the program office at the Space and Naval Warfare Systems Command, PMW-155 (PE-0603207N).


Appendix: Acronyms


ABI – Advanced Baseline Imager
AIRS – Atmospheric Infrared Sounder
AMSR – Advanced Microwave Scanning Radiometer
AMSU – Advanced Microwave Sounding Unit
ASCAT – Advanced SCATerometer
ATMS – Advanced Technology Microwave Sounder
AVHRR – Advanced Very High Resolution Radiometer
CMIS – Conical Microwave Imager Sounder
CRIS – CRoss-track Infrared Sounder
DMSP- Defense Meteorological Satellite Program
FY-2C – Fen Yun
GIFTS – Geostationary Imaging Fourier Transform Spectrometer
GOMS – Geostationary Operational Meteorological System
GPM – Global Precipitation Mission
HES - Hyperspectral Environmental Sounder
HIRS – High resolution Infrared Radiation Sounder
IOMI – Indian Ocean Metoc Imager
METOP – METeorological OPerational (European)
METSAT – Meteorological Satellite
MHS – Multispectral Humidity Sensor
MTSAT – M Transportation Satellite
MODIS – MODerate Resolution Imaging Spectroradiometer
MSG – Meteosat Second Generation
NPOESS - National Polar-orbiting Operational Environmental Satellite System
NPP- NPOESS Preparatory Project
OLS – Operational Linescan System
SeaWiFS – Sea viewing Wide Field Sensor
SSM/I – Special Sensor Microwave/Imager
SSMIS – Special Sensor Microwave Imager/Sounder
SSM/T1/T2 – Special Sensor Microwave Temperature/Humidity Sounder
TRMM – Tropical Rainfall Measuring Mission
TMI – TRMM Microwave Imager
VIIRS- Visible/IR Imager Radiometer Suite



Acknowledgments. The authors thank the following individuals for their contributions to this report: Mark DeMaria, Kurt Brueske, Julie DeMuth, Tim Olander, Harold Pierce, Russ Elsberry, Ray Zehr, Timothy Liu, Joanne Simpson, Bob Simpson, John Knaff, Trisha Gregory and Leanne Avila. This report contains excerpts from the book “COPING WITH HURRICANES -- A Historical Analysis of 20th Century Progress” which has just been published by the American Geophysical Union. We gratefully acknowledge the support of our research sponsors, the Office of Naval Research, Program Element (PE-060243N) and the Oceanographer of the Navy through the program office at the Space and Naval Warfare Systems Command, PMW-155 (PE-0603207N).




References

Bankert, R.L., and P.M. Tag, 2002: An automated method to estimate tropical cyclone intensity using SSM/I imagery. J. Appl. Meteor., 41, 461-472.
Berger, H., and C.S. Velden, 2001: Assimilation of GOES rapid-scan winds into an experimental ETA model during hurricanes. 11th Conf. Satellite Meteor., Madison, WI, 260-261.
Brueske, K. F., and C. S. Velden, 2002: Tropical cyclone intensity estimation using NOAA-KLM Advanced Microwave Sounding Unit (AMSU) data, accepted in Mon. Wea. Rev.
Demuth, J. L., M. DeMaria, J. A. Knaff, and T. H. Vonder Haar, 2000: An objective method for estimating tropical cyclone intensity and structure from NOAA-15 Advanced Microwave Sounder Unit (AMSU) data. Preprints, 24th Conf. Hurr. and Trop. Meteor., May 29 – June 2, 2000, Fort Lauderdale, FL, Amer. Meteor. Soc., 484-485.
Dvorak, V., 1975: Tropical cyclone intensity analysis and forecasting from satellite imagery. Mon. Wea. Rev., 103, 420-430.
Dvorak, V., 1984: Tropical cyclone intensity analysis using satellite data. NOAA Tech. Rep. NESDIS 11, 47 pp. [Available from NOAA/NESDIS, 5200 Auth Rd., Washington, DC 20233.]
Edson, R.T., and J.D. Hawkins, 2000: A comparison of scatterometer-derived wind data over tropical cyclones as determined from ERS-2 and QuikSCAT data. 24th Conf. on Hurricanes and Tropical Meteorology, Amer. Meteor. Soc., 195-196.
Edson, R. T., M. A. Lander, C. E. Cantrell, J. L. Franklin, P. S. Chang and J. D. Hawkins, 2002: Operational use of QuikSCAT over tropical cyclones. AMS 25th Conf. on Hurricanes and Tropical Meteorology, 41-42.
Fu, L., and A. Cazenave, 2001: Satellite altimetry and earth sciences. Academic Press, pp. 463.
Goerss, J. S., C. S. Velden, and J. D. Hawkins, 1998: The impact of multispectral GOES-8 wind information on Atlantic tropical cyclone track forecasts in 1995. Part II: NOGAPS forecasts. Mon. Wea. Rev., 126, 1219-1227.
Hawkins, J. D., T. F. Lee, K. Richardson, C. Sampson, F. J. Turk, and J. E. Kent, 2001: Satellite multi-sensor tropical cyclone structure monitoring. Bull. Amer. Meteor. Soc., 82, 567-578.
Hawkins, J. D., and P. G. Black, 1983: SEASAT scatterometer detection of gale force winds near tropical cyclones. J. Geophys. Res., 88, 1674-1682.
Katsaros, K.B., E. B. Forde, P. Chang, and W.T. Liu, 2001: QuikSCAT facilitates early identification of tropical depressions in 1999 hurricane season. Geophys. Res. Lett., 28, 1043-1046.
Karyampudi, V.M., J.D. Hawkins, E.B. Rodgers, R.M. Zehr, C.S. Velden, J. Simpson, and S. Huntrakul, 1999: Estimating the amount of rainfall associated with tropical cyclones using satellite techniques. WMO Report No. TCP-42, World Meteor. Org., Geneva, Switzerland.
Kidder, S., M. Goldberg, R. Zehr, M. DeMaria, J. F. W. Purdom, C. S. Velden, N. C. Grody, and S. J. Kusselson, 2000: Satellite analysis of tropical cyclones using the Advanced Microwave Sounding Unit (AMSU). Bull. Amer. Meteor. Soc., 81, 1241-1259.
Kummerow, C., J. Simpson, O. Thiele, W. Barnes, A. T. C Chang, E. Stocker, R. F. Adler, A. Hou, R. Kakar, F. Wentz, P. Ashcroft, T. Kozu, Y. Hong, K. Okamoto, T. Iguchi, H. Kuroiwa, E. Im, Z. Haddad, G. Huffman, B. Ferrier, W.S. Olson, E. Zipser, E. A. Smith, T. T. Wilheit, G. North, T. Krishnamurti, and K. Nakamura, 2000: The status of the Tropical Rainfall Measuring Mission (TRMM) after two years in orbit. J. Appl. Meteor., 39, 1965-1982.
Lee, T. F., F. J. Turk, J. D. Hawkins, and K. A. Richardson, 2002: Interpretation of TRMM TMI images of tropical cyclones. Earth Interactions E-Journal, 6, In press.
Liu, W. T., H. Hu, and S. Yueh, 2000: Interplay between wind and rain observed in hurricane Floyd. Eos, Trans. of AGU, 81, 253-257.
Liu, W.T., 2002: Progress in scatterometer application. J. Oceanogr., 58, 121-136.
Rocken, C., R. Anthes, M. Exner, D. Hunt, S. Sokolovskiy, R. Ware, M. Gorbunov, W. Schreiner, D. Feng, B. Herman, Y.-H. Kuo and X. Zou, 1997: Analysis and validation of GPS/MET data in the neutral atmosphere. J. Geophys. Res., 102, 29849-29866.
Shay, L. K., G. J. Goni, and P. G. Black, 2000: Effects of a warm ocean feature on huricane Opal. Mon. Wea. Rev, 128, 1366-1383.
Soden, B. J., C. S. Velden, and R. E. Tuleya, 2001: The impact of satellite winds on experimental GFDL hurricane model forecasts. Mon. Wea. Rev., 129, 835-852.
TAO: Terrestrial, Atmospheric and Oceanic Sciences, 2000: Special issue for applications of the Constellation Observing System for Meteorology, Ionosphere and Climate (COSMIC), 11, 1.
Turk, F.J., G.D. Rohaly, J.D. Hawkins, E.C. Smith, F.S. Marzano, A. Mugnai, and V. Levizzani, 1999: Meteorological applications of precipitation estimation from combined SSM/I, TRMM and infrared geostationary satellite data. In Section 3.2 of Microwave Radiometry and Remote Sensing of the Earth's Surface and Atmosphere, P. Pampaloni Ed., VSP Intern. Sci. Publisher, Utrecht (The Netherlands), ISBN 90-6764-318-4, 353-363.
Velden, C. S., T. L. Olander, and S. Wanzong, 1998a: The impact of multispectral GOES-8 wind information on Atlantic tropical cyclone track forecasts in 1995. Part I: Dataset methodology, description, and case studies. Mon. Wea. Rev., 126, 1202-1218.
Velden, C. S., T. L. Olander, and R. M. Zehr, 1998b: Development of an objective scheme to estimate tropical cyclone intensity from digital geostationary satellite infrared imagery, Wea. and Forecasting, 13, 172-186.