|
||||||||||||||||||||||||||||||||||||||||||
|
Time is a precious commodity when information must be gathered, plotted and presented to the forecaster before warnings can be prepared and distributed, all on a rigid time schedule. Consider the expression below. The time available to a forecaster (A) to assess the latest information relating to a tropical cyclone and make a prediction about its future state can be expressed as follows:
W - O = T + D + A + P, (1)
where:
W = Warning issue time
O = Observation time
T = Time taken for observational data to arrive at TCWC
D = Time taken for data to be processed and then presented to the forecaster
A = Forecaster analysis, assessment and prediction time
P = Time needed for message preparation.
The optimum situation for the forecaster is to have A as large as possible. To achieve this, T, D, and P must be made as small as possible. TCWC managers should be continually examining ways to reduce the time taken for these operations in order to give staff the maximum possible time for assessing the situation and arriving at a sound, reasoned forecast policy.
As an example, synoptic observations made at 0300 UTC can take up to 30 min to arrive at a TCWC (T = 30). In order that the observations are presented to the forecaster, they must be plotted on a chart (unless the forecaster can both remember the codes and visualise its location relative to the cyclone's centre). The plotting can take up to 45 min manually (depending on the volume of traffic) but considerably less by machine. With manual plotting then, the forecaster would generally receive a chart at between 0415 UTC and 0430 UTC, which would allow (for a 0600 UTC warning) around 90 min for assessment of the situation and warning preparation. Direct ingestion of the observations into a workstation would further reduce D.
A geostationary satellite observation taken at 0300 UTC (Observation time) would generally be received at a TCWC at 0330 UTC (T = 30 min). If ingested directly into a workstation the imagery can be analysed by a forecaster almost immediately (D = 0 approx.). However if some manual processing of the data is necessary (photographic printing for example), D may be substantially longer.
In the situation where a radar is collocated within a TCWC or a radar network is directly linked to a TCWC, a radar observation will take a very short time to reach the forecaster and can be processed and presented to make T + D small. Observations can also be received frequently. This would allow a forecaster an almost continuous monitoring of the tropical cyclone for assessment, even during the preparation of the warning message.
Responses from the IWTC-II survey showed that around 55% of TCWCs analysed hand plotted synoptic charts while the remainder had synoptic data machine plotted. Of those offices with machine plotted charts, 30% also had access to synoptic data via a workstation. None of the offices with hand plotted charts used workstations to view data.
The majority of TCWCs then, are in the situation of having lengthy delays while incoming synoptic observations are manually plotted. This practice would tend to decrease the time available for assessing the synoptic situation (large D or data processing time), particularly in relatively data rich areas. Significant economies of time could be made in a number of TCWCs by introducing machine plotting capabilities or by direct ingestion of synoptic data into a workstation (or preferably both).
There are other clear advantages for having machine plotted charts; a machine gives a uniform and overall cleaner presentation, it ensures that all data are plotted, and it is fast.
There are a few problems in introducing machine plotting and these must be recognised early and catered for appropriately. Data must be in the correct format; a machine will not recognise an observation unless it is stringently correct. Thus some human intervention is still necessary for error correction, and machine plotting can never circumvent the late arrival of observations.
All participants surveyed at IWTC-II indicated that their TCWC had access to satellite imagery. Forty two per cent had access to hourly or better satellite data while another 42% could access three hourly imagery from geostationary satellites (compare with McBride and Holland, 1987). At present all of the tropical cyclone basins are under surveillance by geostationary satellites, however 16% of TCWCs do not have direct access to this type of imagery and must rely on polar orbiting satellites to monitor tropical cyclone activity. Several TCWCs, notably Fiji, Mauritius and La Reunion, are not able to monitor the total area of their responsibility with geostationary satellite imagery.
The analysis of geostationary satellite images is the preferred option for monitoring tropical cyclone motion and intensity, as it provides that continuity of observation that cannot be achieved with polar orbiting satellites alone.
Satellite imagery is presented to the forecaster in various ways. Paper copy of satellite imagery is still considered very useful. Of course the better the quality of reproduction the more useful it is. Twenty per cent of TCWCs at present have to work with low quality paper copies, while 70% have access to high quality glossy paper prints. Half of the TCWCs are able to analyse satellite imagery via a personal-computer workstation, which allows the use of more sophisticated analysis techniques. Personal -computer workstations make available useful manipulations of a satellite image, such as digital enhancements, animation (for Geostationary satellites), and channel manipulation (for AVHRR reception).
The IWTC-II survey indicated that 62% of offices possessed image enhancement facilities, 58% could use an image magnification (zooming) function, and 54% were able to loop sequences of images.
One third of TCWCs had none of these facilities available to forecasters. Half of the respondents indicated they had access to satellite derived cloud drift winds at their TCWC.
Recent satellite technology expected to come into increasing operational use in future years is the microwave imager. The U.S. Air Force Special Sensor Microwave/Imager (SSM/I) system is already on board a polar orbiting satellite (Negri et al., 1989). The microwave sensor provides essentially a "smeared" radar image which can help forecasters in determining centre locations of less intense or developing tropical storms which do not display an eye pattern on conventional satellite imagery (Velden et al., 1989). This system also shows promise for improved rainfall estimates (Olson et al., 1989) and surface wind estimates outside of high rainfall rate areas where wind speeds are less than 30 to 50 knots (Rappaport and Black, 1989).
Radar is a powerful detection tool for tropical cyclone forecasters but 25% of TCWCs surveyed did not have access to a radar display. Half of these offices indicated however that they do have access to radar information through coded bulletins. For 30% of respondents, the radar is collocated at their forecasting office, while the remainder accessed imagery from remotely located radar sites. Only 25% of TCWCs have their radar imagery displayed on a workstation.
personal computer technology can import high quality digital radar imagery from remote sites into a TCWC, which can then be manipulated to effectively monitor a tropical cyclone (a good example of this is the Australian RAPIC system).
With a view to future directions for radar in TCWCs, the addition of Doppler capabilities (Doviak and Zrnic, 1984) for the next generation of U.S. National Weather Service (NWS) radar systems, or NEXRAD, will add a new dimension to tropical cyclone detection.
Scheduled for installation in a network along the U.S. Gulf of Mexico and Atlantic coasts during the early to mid-1990s, the Doppler capabilities can provide valuable information on tropical cyclone wind fields and changes in wind fields both prior to and as a storm moves inland (Woods and Marks, 1989), allowing more precise warning advice. Sheets (1990) envisages that while hurricane warnings will be issued similarly to the present, the NEXRAD system will be employed by local NWS offices to provide short-term warnings of the approach of rainbands, destructive winds and possible tornadoes towards specific locations. The NEXRAD system should also assist in improving forecasts of rainfall intensity and so permit better warnings for inland river flooding.
The critical requirements in the presentation of radar data are that the radar receiving equipment in the TCWC needs to be in close proximity to the cyclone forecaster and that it should be an interactive display.
Although the use of reconnaissance aircraft is not widespread through the tropical cyclone basins due to the expense involved in their operation, data from this source were used by around 20% of countries surveyed at IWTC-II. All of these countries were located in the Atlantic basin, and used data from U.S. Air Force reconnaissance and NOAA research aircraft.
Communication of real time data from these aircraft to the US National Hurricane Center was originally made by radio/voice link which significantly limited the volume of information which could be transmitted. However the development of an Aircraft Satellite Data Link (ASDL) (Pifer et al., 1978; Parrish et al., 1984) has permitted much more data to reach the forecaster in real time. Further, it is expected that real time aircraft-based Doppler radar data will be made available to forecasters in the near future. These systems can provide entire data fields within several miles of the path of the aircraft (Jorgensen, 1984; Marks and Houze, 1987).
The expansion of the use of reconnaissance aircraft in other parts of the world can only be achieved if sound economic arguments on the cost effectiveness of this type of operation are submitted to and accepted by governments. In the harsh economic reality of tight budgetary restraint on government spending, this is a daunting task; particularly so because, while the cost of new resources is overt and quantifiable, the benefits they bring are predominantly covert, perspective dependent, and difficult to measure (Woodcock, 1989).
An alternate approach has been proposed which advocates the use of long endurance, remotely-piloted aircraft to observe tropical cyclones. The Perseus (Emanuel and Anderson, 1991) has been designed to fly at altitudes up to 18 km and to periodically drop sondes (up to 100 in the payload) into the tropical cyclone. The Autonomous Aerosonde (Holland et al, 1992) is a smaller aircraft weighing less than 20 kg, and is planned to carry on-board meteorological sensors to provide radiosonde-quality observations as it orbits within the eye region of the storm. The emerging technology of light-weight computing, communications and navigation electronics makes these cost-effective observing systems a promising source of information in the future.
In recent years there have been significant advances in the development of 3-dimensional numerical models. Not only are limited area models (LAMs) being used to make real time predictions of tropical cyclone motion but also global models. Major current models that are used for tropical cyclone forecasting are described in other sections.
Numerical model output is becoming an increasingly reliable source of guidance for operational forecasters. Even at present, 50% of countries surveyed at IWTC-II indicated having access to guidance from global models. This group included smaller countries with little computing capability of their own. Advice issued by the U.K. Meteorological Office (UKMO) is sent routinely twice a day to provide general forecasting guidance on tropical cyclone motion. Morris and Hall (1989), showed aspects of the performance of the UKMO model with regard to tropical cyclones (with maximum winds in excess of 50 knots) in the North Atlantic and North Pacific Oceans. The study indicated most skill from the model in the 2 to 5 day period.
The use of numerical guidance from ECMWF for tropical cyclone motion has been discussed by Reed et al. (1988) and Chan and Lam (1989). Operationally, forecast positions can be plotted on local area charts, primarily to give forecasters basic guidance on track movement (Fig. 6.2). This procedure can also give the forecaster a check on the consistency of the model output from run to run, as well as allowing an intercomparison of trends between different models (including statistical/analogue and conventional techniques).
Figure 6.2: Track forecasts (January 1990) plotted from the UKMO advisories for general guidance. Such guidance allows smaller countries access to global numerical products.
When several different forecasts are available, consideration should be given to implementing approaches such as that of Leslie and Fraedrich (1990), who suggest position errors can be reduced by a combination of independent forecasts.
The ingestion of information from numerical guidance sources into a TCWC can be made more efficiently through the use of a workstation, as the raw transmission data can be quickly processed into a useable form.
| Bureau Home || BMRC Home || Search || Contact BMRC Webmaster |
Home | About Us | Learn about Meteorology | Contacts | Search | Help | Feedback Weather and Warnings | Climate | Hydrology | Numerical Prediction | About Services | Registered Users | SILO |
|
© Copyright Commonwealth of Australia 2008, Bureau of Meteorology (ABN 92 637 533 532) Please note the Copyright Notice and Disclaimer statements relating to the use of the information on this site and our site Privacy and Accessibility statements. Users of these web pages are deemed to have read and accepted the conditions described in the Copyright, Disclaimer, and Privacy statements. Please also note the Acknowledgement notice relating to the use of information on this site. No unsolicited commercial email. |
