These pages were created in 1999-2000, being based on research which was carried out from mid-1980's till the end of 1990's. Though we seem to know much more nowadays (2011) about storm-top processes and some of the observed features, many questions still remain open. We keep the pages below namely for historical purposes, as an archive of cases which were at the beginnings of storm-top 3.5-4.0 Ám reflectivity and above-anvil plumes research.
2011, on the occasion of migration of these pages to a new location, I
removed some of the never-finished parts or pages, and added a recent
list of publications, documenting the progress of this research in the
years which followed after 2000.
Finally, I want to thank to all of those who contributed significantly to this early research, my colleagues and friends (in the order as we met for the first time): Vincenzo Levizzani, Chuck Doswell, Pao Wang and Bob Rabin. And also to my CHMI colleagues, Karel Hlavaty and Pavel Hampl, both of them helping me crucially with the early computations and data processing.
The original pages follow.
Convective storms, one of the nature's most violent (and at the same time one of the most spectacular) phenomena, are typically studied by means of weather radars, lightning detection networks or surface observations. Greatest advantage of weather radars is in their capability to see "what's inside", what is the 3D structure of a storm (at least in terms of radar observations: radar reflectivity, Doppler wind fields, polarimetric radar data, etc.). Lightning detection networks provide us with information about storms' electric activity - e.g. polarity, intensity and location of discharges and their frequency. Surface observations inform about accompanying weather - wind speed, precipitation intensity, hail occurrence, tornadoes and microbursts… Radar and lightning detection networks are nowadays one of the basic tools for nowcasting (detection and very short term forecasting) of severe storms.
The task of first weather satellites was to bring us information about phenomena on a much larger scale than that of convective storms - frontal systems and cyclones. However, as the quality of satellite sensors has improved and namely with launch of the first geostationary satellites, convective storms soon became one of the targets of researchers dealing with meteorological satellite data. Since the images "from above" display only tops of convective storms, it is not as easy to infer from these about storm severity or their 3D structure. Therefore a great effort has been put namely on studies of links between storms' cloud top thermal IR field and their structure or severity. This has revealed certain thermal IR cloud top features, typical for severe storms - e.g. "cold-U" (or "cold-V") and embedded warm spots inside them. Similar research has been done in visible and near IR bands, but with not as clear or promising results.
Much less attention has been devoted so far to spectral bands situated approximately at 3.5 - 4.0 Ám. These bands are the only ones where the radiance detected by satellite results from two different origins, from two different components - emitted and reflected ones. This makes the interpretation of images in this band somewhat more difficult, less straightforward. However, on the other hand it provides us with information about microphysical composition of cloud tops, which makes this band very attractive. The appearance of convective storms in this spectral band - in context with images from other wavelengths - is one of the main topics of these pages.
You will not find too much of "deep science" here - only the necessary information to be able to "read" the images, to understand what you are looking at. The main goal of these pages is to present a "gallery" of the most interesting cases, to show the storms with "increased 3.7/3.9 Ám reflectivity" or with "plumes" in all their broad variety (these terms are explained and discussed in following sections). Typically, the space in journals or conference proceedings is rather limited and not too many examples can be shown there. Moreover, animations (loops) showing storm development or fast switching between different spectral bands are impossible in printed form. These limitations of printed form were among the main impulses for creation of the pages you are just going to read or browse.
The structure of these pages more or less follows the history of this research. The first section - Storms in the 3.7 Ám spectral band
- deals in general with appearance of storms in that band, bringing you
some of the older, though still very interesting cases of the
"increased 3.7 Ám reflectivity" of storm tops (this section is
based on the NOAA/AVHRR data). Also, a method for extraction of
the reflected component from this band and computation of the
3.7 Ám reflectivity is discussed here. Plumes above storm tops
can be one of the forms of the increased 3.7 Ám reflectivity.
Though originally revealed in the AVHRR channel 3 of the NOAA
polar orbiting satellites, some of them can be found also in other
spectral bands or in data from other satellites (Meteosat, GOES). Next folowed a joint U.S. - Czechoslovak project nicknamed "MOST" (1994-1997). Since MOST stands for Multispectral Observations of Storm Tops,
you can easily guess what was the target of that project. Storm tops,
as seen by NOAA polar orbiters as well as by GOES-8/9 satellites, have
been examined in conjunction with Doppler weather radar observations.
Though MOST did not answer to many of our questions, it certainly
brought new insight into this research. Some of the MOST results are described in the referenced papers.
These pages in no case attempt to be a tutorial about weather satellites and their sensors, multispectral image interpretation, nor a textbook about convective storms - you will find that information elsewhere (see our References and related papers page). Where possible, the additional information is "linked" right in the text.
Many of the older cases are archived in a data format which we are not able to process easily anymore (namely remap the original data into a common projection format or calibrate these), thus it has been somewhat difficult task to decide about how to show clearly the location of a particular storm and the image scale. Where possible, the "location" button brings up an image with national borders and coastlines (in polar stereographic projection), while elsewhere you will find at least a geometrically rectified image showing larger part of Europe, with no borders and no coastlines - hopefully you can find some well known landmarks there (sorry for this inconvenience)... In all of these "location" images, the storm shown next in detail is marked by a red arrow. An approximate scale of the detailed images can be found in the status bar of your browser (at the bottom of your browser window).
The enhancement of images is not "standardized" - it may be somewhat different from case to case. The composite images of AVHRR channels 1, 2 and 4 show highest cloud layers in bluish or white shades, while low and middle clouds are rather yellowish. Images in the AVHRR channel 2 are enhanced to show as much as possible details in cloud top morphology. Thermal IR images - the AVHRR channel 4 - are enhanced to show details of the brightness temperature field; red color depicts the coldest tops of storms (typically the "overshooting tops"), and blue to violet shows the warmer anvil regions or warmer clouds. Similar enhancement is used also for Meteosat or GOES images in corresponding bands. Visualization and enhancement of the 3.7/3.9 Ám images are discussed in the next section, devoted to this spectral band.