Page Contents:
Tropical rainfall forecasting is a
challenging task. It is common that rainfall forecasting for day one is reliable,
but that forecasting for the second, third and more distant days in time, forecast reliability, which is a
quantitative measurement of similarity between forecasts and observations, will
degrade significantly. To improve tropical rainfall forecasting is important,
particularly when it comes to hurricane tracks and rainfall accumulations.
Moreover, the improvement will have an integrated effect on the model global scale
behavior of rainfall, and help us understand the model rainfall patterns.
In order to improve fainfall forecasting, atmospheric scientists have
concentrated on three main aspects: improving numerical weather forecast models; acquiring accurate observations for model comparison and validation; and developing analysis or initialization methods, which incorporate observational
data into the model and give accurate initial conditions for the forecast.
One major accomplishment deriving from the use of TRMM satellite data is
shown by the short-term rainfall forecast study carried out at Florida State
University. In this study, TRMM rainfall data and SSMI data are included in
the multi-analysis process of the numerical weather forecast model. The
results show that global, as well as regional forecasting reliability are
enhanced by the addition of the TRMM data to the data pool. TRMM data are also
used as the validation data for deriving the statistical parameters of the
multi-model super ensemble system. The resulting forecast correctly predicted
the tracks of major hurricanes in 1999 as well as dramatically increasing the
accuracy of rainfall forecasting. Scientists attribute this success to a
combination of improved analyses delivered by the super-ensemble approach and
inclusion of accurate rainfall estimates over the tropics from the TRMM
satellite.
The figures below exhibit a three-day rainfall forecast made by the
multi-model super ensemble method.
The top panel shows the initial precipitation field of the
forecast model. The initialization process of the model makes use of
the precipitation observed by TRMM TMI and SSMI (Special Sensor Microwave Imager) at 12Z Oct. 27, 1999. The heavy precipitation (red shading) along with
supercyclone Orissa occurs over South Asia, over Bay of Bengal, and around
Indonesia. The following three pairs of figures show the comparison between
the forecast and the observed precipitation for the next three days. For day one,
the forecast rainfall field seems identical to the observed in terms of
rainfall regions and intensity. For the day two and day three, the predicted
precipitation are still fairly close to the observation, but the discrepancies
seem to develop around Orissa's center precipitation region.
The researchers at Florida State University currently extend their
short-term rainfall forecast to thirty-day forecast experiment using TRMM and
SSM/I data in near real time. For the detailed description about this study,
please see TRMM
site , then link to "Latest News",
see released news on 01-12-2000 . The detailed
description of multi-model and multi-analysis forecast model are given by
Simpson et al. and by Krishnamurti et al.
References
Simpson, J., Kummerow, C.D., Meneghini, R., Hou, A., Adler
, R.F., Huffman, G., Barkstrom, B., Wielicki, B., Goodman, S.J., Christian, H.,
Kozu, T., Krishnamurti, T.N., Yang, S., Ferrier, B., 2000: The tropical rainfall
measuring Mission (TRMM) progress Report, Earth Observation and Remote Sensing, Vol. 18, August Issue.
Krishnamurti, T.N., Kishtawal, C.M., Bachiochi, David, Zhang, Zhan, Larow, Timothy, Williford, Eric, Gadgil, Sulochana, and surendran, Sajani, 1999: Multi-Model Superensemble Forecasts for Weather and Seasonal Climate. Submitted to
Meteorol. Atmos. Phys.
TRMM 2A12 are derived from TMI data. The data description and figures are
provided on Cloud and
Precipitation Formation topic of this page.
|
 |
Atmos Composition
