Скачать презентацию Toward a Global Climatology of Tropical Cloud Clusters Скачать презентацию Toward a Global Climatology of Tropical Cloud Clusters


  • Количество слайдов: 1

Toward a Global Climatology of Tropical Cloud Clusters Christopher C. Hennon and Charles N. Toward a Global Climatology of Tropical Cloud Clusters Christopher C. Hennon and Charles N. Helms University of North Carolina Asheville, Asheville NC Introduction Kenneth R. Knapp NOAA/National Climatic Data Center, Asheville NC Figure 1. Geostationary IR imagery of the tropics for August 9 th 1999 at 21 UTC. Cloud clusters are circled in red. Variable Latitude/Longitude Weighted Lat/Lon Pixel Count Mean Tb Minimum Tb Median Tb Standard Deviation of Tb Coldest 5 th Percentile Tb Coldest 10 th Percentile Tb Maximum Radius Tropical cyclones develop from areas of localized intense convection known as “cloud clusters. ” Although the process by which cyclogenesis occurs is an important area of research, progress is hampered by the lack of any long-term database of these systems. Several tropical cyclogenesis studies (e. g. Hennon and Hobgood 2003) describe the large effort required to create cloud cluster datasets that cover only a few years. Presented here is an automated method of identification based solely on geostationary infrared satellite imagery, with the aim of creating such a dataset. Data The gridded ISCCP B 1 dataset (Knapp 2008) provides the algorithm with global infrared satellite imagery coverage from 1980 through 2008 (see Figure 2). These data are available in 3 -hour increments at 8 km resolution. Since our goal is to identify “seedlings” rather than mature storms, we remove developed TCs from the dataset. We use the International Best Track Archive for Climate Stewardship (IBTr. ACS, Knapp et al. 2010) dataset to identify and remove developed clusters. IBTr. ACS contains best track data from every forecast and analysis center that produces them. Figure 3. Global track map for 1999. Developing clusters are plotted in red, nondeveloping in blue, and clusters occurring outside of their basin’s active season in grey. Description Units Coordinates of geometric center degrees North/East Coordinates of max. convection degrees North/East Num. of pixels within threshold Average brightness temperature K Coldest brightness temperature K Median brightness temperature K Std. deviation of brightness temp. K 5% pixels colder than this K 10% pixels colder than this K Largest distance around azimuth from km center to edge of cluster Minimum Radius Smallest distance around azimuth from km center to edge Mean Weighted Radius Average distance around azimuth from km center to edge Maximum Cloud Top Height Tallest cloud top height km Mean Cloud Top Height Average cloud top height km Translation Direction of movement compass degrees Translation Speed Velocity kt Quality Control Flags Identifies developing CCs, times that are interpolated, and general quality Table 2. Summary of several key output variables. Figure 2. Timeline of ISCCP data coverage. Methodology Cloud clusters are identified using five parameters: brightness temperature, size, persistence, independence, and location. The individual requirements are listed in Table 1. The area of intense convection is defined here by a brightness temperature threshold, which varies by ocean basin. Developed clusters are identified based on the mean distance to a tropical cyclone position in IBTr. ACS. If a match is found, the genesis location is recorded in the dataset and a flag is set to identify the cluster as “developing”. No further coordinates are recorded for that system. Once a cluster is identified, it is tracked until it moves over land or has dissipated (fails one or more requirements). To account for the irregular convection in these systems, clusters are allowed to fail one or more of the requirements for up to 12 hours before they are considered to have dissipated. Positions are linearly interpolated in these instances, but cluster variables are not calculated. Parameter Results The database will cover the period from 1980 through 2008 and include statistics on location, genesis information, cloud top brightness temperatures, size, and cloud top height. A summary of the key output statistics is given in Table 2. Due to inconsistent satellite coverage, cloud cluster numbers are artificially low prior to 1982. The impact of this coverage can be seen in Figure 5. The 27 - year span of cloud clusters approaches climatological averaging periods and could be a useful resource for studying the impact of climate change on the tropics. Figures 1 and 6 show examples of identified cloud clusters on an infrared image. The global cloud cluster tracks for 1999 are plotted in Figure 3. For an example of a single cloud cluster track from the same year, see Figure 4. Radius of intense convection > 1° Equivalent cloud cover > 90% of 1° radius circle Persistence Latitude < 30°, over water Table 1. Parameters used to identify cloud clusters Additional metadata, such as NCEP/NCAR reanalysis variables, will also be included in future releases. Statistics for cluster speed and direction are currently in development and may be included in the first release. Other areas of future development include improving the tracking accuracy and removing non-cluster convection. Acknowledgments This work is funded by the NOAA Climate Change Data and Detection (CCDD) program, opportunity No. OAR-CPO-2009 -2001430. Special thanks to Amanda Bowen, who performed the threshold temperature calculations, and the rest of the IBTr. ACS team at the National Climatic Data Center who provided suggestions and guidance. We would also like to recognize Mark De. Maria, who initially suggested that this type of product would be very useful. Knapp, K. R. , M. C. Kruk, D. H. Levinson, H. J. Diamond, and C. J. Neumann, 2010: The International Best Track Archive for Climate Stewardship (IBTr. ACS). Bull. Amer. Meteor. Soc. , 91, 363 -376. Distance to other clusters > 1200 km Location The first version of the cloud cluster data is currently going through quality control and is expected to be released by the end of 2010. Plans for future development include an operational version aimed at supporting cyclogenesis forecast models. Hennon, C. C. and J. S. Hobgood, 2003: Forecasting tropical cyclogenesis over the Atlantic basin using large-scale data. Mon. Wea. Rev. , 131, 2927 -2940. Cluster lifespan > 24 hours Independence Future Work References Below threshold temperature (basin dependant) Size Figure 5. Total and developing clusters identified each year. Requirement Brightness Temperature Figure 4. Cluster track from 1999. The cluster lasted from 12 UTC 9 July through 21 UTC 11 July. The “*” indicates the initial position and the “+” indicates the final position of the cluster. During the entire 1980 -2008 period, the algorithm identifies a total of 45, 708 cloud clusters. Of these, 2, 362, or ~5%, are developing cases. The number of identified developing clusters is approximately 75% of the total number of developing cyclones existing in IBTr. ACS, suggesting that some TCs may be started without meeting our identification requirements for convection. For 1982 -2008 period, an average of 1, 611 cloud clusters are identified annually; the average number of developing cases each year is 83. Figure 6. Global IR imagery for August 9 th 1999 at 21 UTC. Cloud clusters are circled in red. Knapp, K. R. , 2008: Scientific data stewardship of International Satellite Cloud Climatology Project B 1 global geostationary observations. J. of Applied Remote Sensing, 2, 023548.