-- Mitchell K. Hobish (mkh@sciential.com), reprinted from Earth Sciences News, Vol. 3, No. 1
It's always a welcome situation when a recently launched satellite begins to transmit the data it was sent to acquire. The members of the Tropical Rainfall Measuring Mission (TRMM) science and engineering teams were elated when several of the instruments on TRMM recently demonstrated the robustness of their design by providing data thatin the case of the Precipitation Radareven exceeded the science team's expectations.
The key objectives of the TRMM program are to obtain and study multiyear science data sets of tropical and subtropical rainfall measurements; understand how interactions among the sea, air, and land surfaces produce changes in global rainfall and climate; help improve modeling of tropical rainfall processes and their influence on global circulation in order to predict rainfall and its variability at various time-scale intervals; and test, evaluate, and improve the performance of satellite rainfall measurement techniques.
The primary scientific instruments on TRMM include the TRMM Microwave Imager (TMI), the Visible Infrared Scanner (VIRS), and the Precipitation Radar (PR). These three instruments work in a complementary manner, and serve as a kind of "flying rain gauge." The VIRS will bridge the gap from the TRMM rainfall estimates to those estimates made with geostationary and lowEarth-orbiting satellites using visible/infrared techniques. This will help overcome the temporal and spatial sampling problem inherent in a single satellite's coverage. This primary complement is accompanied by the secondary scientific instruments, a Clouds and the Earth's Radiant Energy System (CERES) and a Lightning Imaging Sensor (LIS).
The TRMM observatory was successfully launched on November 27, 1997 from Japan's Tanegashima Space Center by a Model HII launch vehicle. The 96-minute orbit is inclined 35o to the Equator and is circular, with an altitude of 218 nautical miles (350 km). This orbit allows collection of tropical rainfall data from the passive microwave sensor (TMI) every 16 orbits, with a diurnal precession rate arranged so that it comes over any given point roughly an hour later each day, so it repeats approximately every 30 days. The orbit was intentionally kept low in order to resolve radiances over the 10 km scales typical of convective clouds.
"It's really too early to draw conclusions about total rainfall estimates," says Chris Kummerow, the TRMM Project Scientist, "but indications are that the data coming down from the recently calibrated sensors will make significant contributions to our understanding of rainfall mechanisms and processes." That's not to say that the preliminary data aren't useful; quite the contrary.
Indeed, Kummerow noted that, "With TRMM's high-resolution radar data, passive microwave data, infrared (IR) data, and lightning data, we are finding unambiguous evidence that the passive microwave is consistently putting more rainfall over the oceans than the IR, and consistently less rainfall over the land than the IR. The lightning data show that lightning is confined almost strictly to over-land cases, with very little lightning over the oceans."
He continued, "The nice thing about TRMM is that we can subset all the different data sets to exactly the same time and area. We've known for a while that the physics from IR data from geostationary sensors wasn't as good as that from passive microwave. But, by virtue of looking at a scene basically all the time, it does provide better sampling. But we could never untangle how much of what we saw was due to sampling, and what was due to errors. So, by being able to subset everything to the same sampling, we can really start looking at the physics."
Such insights allow the TRMM team to explore the differences in ice scattering between over-land storms and oceanic storms. Said Kummerow, "Ice scattering at the shortest wavelength of the passive microwave sensor appears to be about the same in these storms. That, we think, is indicative of the total amount of ice being roughly equivalent in oceanic and land storms. But, when we start going to longer wavelengths, the ocean storms don't show any signature. We think that's because the ice crystals are smaller in the oceanic storms."
Such surmises would acquire increased validity with more thorough analysis and with collection of data from aircraft flights through the storms synchronized with TRMM passes. In conjunction with ground-based radar studiesthe aircraft studies are to begin in April in Houston, then move to Florida in the summer, Brazil in January 1999, and the Kwajalein Atoll in September 1999these data will allow TRMM scientists to determine just what is happening in these storms.
The TRMM science team already has had a few surprises, requiring some fast reworking and development of algorithms to analyze the data. Initial radar data analysis has given indications that estimates of radiative transfer to space may be in error. According to Kummerow, this could come from the models' not putting enough liquid into the atmosphere to match the passive microwave observations. One possible source of error is the drop-size distribution. Kummerow said that, "The raindrops don't all come down in the same size; there's always a distribution of drops, and it's the mean of the distribution that affects the radar. It's clear from these early experiments that something we're doing is not quite right. We may be assuming that the drops are too big, but the most likely mistake is that we're assuming that the drop-size distributions are the same from the surface all the way up."
While drop sizes have been measured near the surface, it is very difficult to get drop-size distributions above the surface. Kummerow says that, to a first-order approximation, this may be the source of the error. The mechanisms of drop-size growth or diminution may be critical to understanding several phenomena. Drop size in stratiform storms, which generate relatively gentle rain, likely diminishes due to evaporation as the drops travel through the atmosphere. On the other hand, in convective storms the raindrops may continue to grow through accretion as they descend.
"None of that has really been incorporated into rainfall algorithms. It appears as though it could have a fairly big impact," said Kummerow. But now, with TRMM data covering footprints much smaller than those used previously, details should be forthcoming. He continued, "The tropical rain systems are mostly driven by sporadic convectionthunderstorm convection. A lot of the mid-latitude rain falls from frontal bands. There's every reason to expect that these two have very different dynamics and have very different features in them. We're very lucky in that TRMM goes as high as 35o latitude. At 35o we do catch a lot of wintertime frontal precipitation."
Several other research and operational areas have been affected by the preliminary TRMM data.
For example, TRMM has pointed up some difficulties in interpreting rainfall echoes from ground-based radar data. According to Kummerow, "When you get sufficiently far away from ground-based sites, ground-based radars can start intercepting the region where the snowflakes melt, giving you very, very large echoes that are not really surface rainfall. This problem is well known to the radar meteorology community, but it's a huge problem when you start providing operational accumulations that are then used drive models in terms of surface moisture. TRMM, even with only occasional overpasses, may be able to sort out some of these ambiguities."
Another area opening up for discussion deals with the processes that give rise to so-called "warm rains," such as are found in Hawaii. In Hawaii, these rains are orographically induced, i.e., when very moist Pacific air runs into the mountains it's lifted up a little, and water condenses outdramatically. Hawaii is the rainiest place in the world, and yet these rains have no signature in the IR.
"Most IR algorithms don't assign rain until you get to about 235 K," said Kummerow, "so we're talking about 40o below zero, Celsius. So, when you have these clouds that never reach freezing level, the IR says (by and large) that there's no rain." The passive microwave sensors over land cannot see anything but ice particles in the clouds, and raindrops over land are indistinguishable from the land itself. "So the passive microwave sensors have been happily assigning a zero rain rate to the rainiest place in the world," says Kummerow. He continued, "With the TRMM precipitation radar we're going to start looking at how much rain we're really missing. We also need to know in how many places this occurs. We know about Hawaii; there are places in India that favor this kind of precipitation. But we don't really know how important it is to global processes: is this a big chunk of rain that we're missing, or is it just a scientific curiosity?"
The answer to this question is important because the total amount of energy that is produced in the atmosphere and drives the worldwide atmospheric circulation is directly proportional to the total amount of rain that falls.
Another area that is being examined is the effect of the diurnal cycle on rainfall. Climate modelers need to know that they are capturing the diurnal cycle properly, because this will tell them if their models properly drive the convection.
"Nagging questions about the accuracy of the physics in the models, the amount of warm rain, and the effects of the diurnal cycle have left us with an uncertainty in global rainfall of about 30%," says Kummerow. "The TRMM observations and the physics we are starting to understand will have a huge impact on our ability to quantify rainfall processes. Errors smaller than 10% are certainly achievable using all the sensors that we have today and will even improve our estimates from previous years as well as into the future."
The data were closely held until mid-May to allow the TRMM team to ensure that the data they released to the larger science community was of decent quality, and that it was formatted properly. They see this as but another service to the community, part of their mandate to provide TRMM data, and to help us understand how rainfall affectsand is affected byour environment.