GISdevelopment.net ---> AARS ---> ACRS 2000 ---> Poster Session 2



The Study of Precipitation Effects on AMSU and Application of AMSU on Typhoon Monotorning

Kung-Hwa Wang
Meteorological Satellite Center, Central Weather Bureau
64, Kung Yuen Rd. Taipei,100
Tel:(2)2349-1238 Fax: (2)2349-1259
E-mail: peter@msc.cwb .gov.tw
TAIWAN

Keywords: AMSU, microwave, precipitation typhoon

Abstract:
AMSU (Advanced Microwave Sounding Units) is a new instrument on NOAA series. A forward model was developed and it included emission and scatter mode. From simulation of this radiation transfer model many characters of AMSU are understood. Total water vapor (TWV) retrieval algorithm was improved and validated with RAOB data. Improved results from estimated TWV compared to RAOB data are clear. AMSU Brightness temperatures (TB) have some relationship with rain rate. There is almost not any evidence shows that the relationship is existing among observed AMSU data and rain rates. AMSU is a very good tool to make typhoon structure more clear, and it is easier to identify center of typhoon than before.

1. Introduction
After NOAA-15 is operating from 1998, AMSU data is the most interesting data in ATOVS data sets. Because it was a new instrument in TIROS-N series, not so many utilization reports were published. Pre-study and research were performing in NESDIS/NOAA and relative association before satellite was lunched. The advantage of AMSU has not effectes by not rain cloud. The specification of AMSU was published in NESDIS User's Guide (Goodrum etc, 1997). Norman Grody (1999) described some useful AMSU retrieval algorithms. Which include TWV, rain identification, scattering index, and Liquid Water Contents (LWC) and it is worldwide use.

AMSU-A is a multi-channel microwave radiometer that is be used for measuring global atmospheric temperature profiles and provide information on atmospheric water in all of its forms (with the exception of small ice particles, which are transparent at microwave frequencies) from the NOAA 15 spacecraft and following series. AMSU-A is a cross-track, line-scanned instrument designed to measure scene radiance in 15 discrete frequency channels which the calculation of the vertical temperature profile from about 3 millibars (45 Km) pressure height to the Earth's surface.

AMSU-B is a 5 channels microwave radiometer. The purpose of the instrument is to receive and measure radiation from a number of different layers of the atmosphere in order to obtain global data on humidity profiles. It works in conjunction with the AMSU-A instruments to provide a 20 channel's microwave radiometer. AMSU-B scan patters and geometric resolution translate to a 16.3 Km diameter cell at nadir at a normal altitude of 850 Km.

2. Development of AMSU forward model

2.1 The background of moded's development
RTTOV-7 was released recently, it seems is the best AMSU forward model. On 1999 JAN. , NOAA-15 is running on orbit for a while and the data is routing receiving. Before using AMSU data, the characteristics of AMSU should be make clear. Except FASCOD3P (Fast Atmosphere CODE version 3 Plus) was in hand, no other's model could be used. FASCOD3P is a line by line model and very precise model, because it is so complicate that uneasy to apply on real data and it is uneasy to modify the codes to fit the requirement of research. A radiation transfer model, which combines emission and shattering modes in microwave bands, is needed. Develop a forward model by myself is the last choice at that circumstance.

In this forward model, which the absorption coefficients of Oxygen, water vapor, ozone and liquid water are adapted from Liebe (1985) and Rosenkranz (1998). Scattering model is used Eddington Approximation, and the coding was developed by Kumerrow (1994). In cloudy and raining condition, the most uncertainty factor is cloud physical, which include cloud height, vertical distribution of liquid water contents, water vapor, rain and ice. Those cloud parameters usually can not be obtained by observe system routinely. CIMSS provided a cloud model in this model, which is a simulation model from hurricane Bonnie (98238).

In order to demonstrate the effects of precipitation on AMSU upward radiance, the parameters of cloud model, which includes water vapor, liquid water content (LWC) and precipitation was normalized as function of surface rain rate. Dependding on surface rain rates and cloud top temperature, cloud types were classified into 9 categories. Each of these 9 categories has different vertical distribution of hydrogen. For different absorption line coefficient there are a little different transmittance in microwave frequencies. About liquid water absorption line, because of coefficients of Kummerrow's are divided into two groups, there is a gap at 100GHz. So Leibe's coefficient is used in this model.

2.2 Validation of the model
A satellite pass in 15-NOV-1998 was chosen as real data. All RAOB data under satellite pass were collected and treated as model's input parameters. Assumed all data set were under clear condition, and surface emissivity was set to same value over ocean, and AMSU channel 5 was chosen for comparison because channel 5 is less effect by surface situation. The compared result of estimated TB and satellite observed TB is shown on Fig. 1. Most of the data is distributed along a line, but some of the point scatter from the strength line, which are caused by precipitation and surface emissivity. This model is can be used for AMSU data simulation.

2.3 Characteristics of AMSU data on precipitation and liquid water content
With different rain rate and maintain other parameters in same situation, then we may get the variance of TB that has effectes by rain rate. AMSU channel 1-4 and channel 15-17 are change significant with rain rate from 0 to about 30mm/hr. AMSU channel 1,2 and 15,16, 17, those are surface channel, and has effecte by precipitation. The effects of LWC on AMSU upward radiance are similar features as effect of precipitation. It is found that the influences of precipitation and LWC have similar intensity on AMSU.

Having 2343 Km swath width, local zenith scan angle of AMSU is up to 57 degree, so the limbs correct is needed for further retrieval process. From the simulation of the AMSU forward model, TB of AMSU channel 1 and 2 are increasing when local zenith angle is increasing, others are decreasing. Sea surface is cold to AMSU channel 1 and 2, surface component is decreasing when local zenith scan angle is increasing, then upward TB of channel 1 and 2 are increasing. TB of other channels decreasing with scan angle increasing is caused by optical depth is increasing.

3. Improvement of total water vapor retrieval
After AMSU data was available, research group of Dr Grody (1999) announced a few algorithms for AMSU instrument to retrieve sea ice, water vapor, cloud liquid water (LWC), rain identification and snow cover. Those algorithms have great contribution on global hydrogen retrieval by AMSU in preliminary stage. The accuracy of TWV could not fit the requirement of local weather services, such as in Taiwan area and tropical area. In this situation, an improved algorithm of TWV retrieval was developed.

In order to reduce the influence of surface emissivity to TB of AMSU, nine RAOB data on island and near Taiwan area were selected. TWV is estimated from those RAOB data and it was regarded as ground truth. Most of the water vapor is located under 3000 meter, so low level WV is very close to TWV. When satellite observed AMSU data pass over those islands RAOB data were collected as well. From the scattering diagrams between TWV and TB of each AMSU channel, AMSU channel 1, 2 and 16 have better correlation with TWV. So channel 1, 2, and 16 of AMSU were chosen for retrieval. Linear multiple regression function was estimated from above TB of channels and collocated RAOB data. Then TWV can be retrieved from AMSU data set. Estimated results are shown on Fig. 2. 3. Statistically it is shown that this improved method is obviously better than original one and the spatial resolution is increased either.

4. Monitoring Typhoon by AMSU
When typhoon is severe, there is almost not problem to identify the center of typhoon. When typhoon or tropical cyclone is weak, or the eye of typhoon is masked by cirrus cloud, it is very difficult to identify the center of cyclone. Traditionally base on the cloud bands of the cyclone to track the eye of typhoon, such as Dvorak (1982). This identification skill strongly depends on the experience of the weather analyst. When center of typhoon is unclear, the forecast of moving direction of typhoon could be missed. Identify center of typhoon is a very important issue in tropical and subtropical area weather services.

Viewing from vertical distribution of AMSU weighting function, each channel of AMSU-A can present vertical structure of temperature at specific level. It is shows in Fig. 4. Viewing the vertical structure over a cross line that passes through center of typhoon. The surface warm core of typhoon and cold core at high altitude is significant. Weak typhoon is shows on Fig. 4. The low-level warm core still is very easy to be identified. Not only center of typhoon, but also the strength of typhoon could be defined also. Further more researches are keep on going.

5. Conclusion
A developed forward model is presented that could be used in real data simulation and widely used in research. Not only clear sky but also cloudy or rain could be a simulation tool also. Using AMSU data on weather analysis was mentioned in this research, such as using AMSU may provide total water vapor or low level water vapor and typhoon tracking successfully. The relation between AMSU and wind speed is an interesting and useful article in weather forecast, and it will be next research object. The original design purpose of AMSU is not used for precipitation retrieval. Actually, the correlation between rain rate and AMSU observed TB is poor, but rains still affect upward radiance of AMSU surface channel. Removing rain effects on AMSU data utilization, or getting more precise surface emissivity to retrieval rainfall information from AMSU is another task in the near future.

Since AMSU data was available two years passed, and NOAA-16 was lunched also. AMSU is not a new instrument any more, and other new microwave instruments are in progressing, which means microwave remote sensing may create a new window for satellite weather service.

Reference
  • Dvorak, V.F., 1982: A technique for the analysis and forecasting of tropical cyclone intensities from satellite pictures. NOAA Tech. Memo. NESS 36, Dept of Comm.
  • Goodrum, Geoffrey, Katherine B. Kidwell, Wayne Winston, 1997, NOAA KLM USER's GUIDE, NOAA/NESDIS
  • Grody, Norman, Fuzhong Weng and Ralph Ferraro, 1999, Technical Proceedings of The Tenth International ATOVS Study Conference, Boulder CO
  • Kummerow, C. L. Giglo,1994: A passive microwave technique for estimating rainfall and vertical structure information from space, Part I: algorithm description, J. of Appl. Meteo. , 33, 3-18.
  • Liebe, H. J. , 1985, An updated model for millimeter wave propagation in moist air, Radio Science, 20, 1069-1089.
  • Rosenkranz, P. W., 1998, Improved rapid transmittance algorithm for microwave sounding channels, International Geoscience and Remote Sensing Symposium (IGARSS'98), Seattle, WA


Figure 1. Scattering diagram of estimated TB and satellite observed TB. Data are almost linear distribute along a line. Some points disperse from central line are caused by unknown precipitaion and surface emissivity.





Figure 2. Scatter diagram of two result of Grody's and Wang's TWV with RAOB's TWV. Lower graphic is the difference between these two algorithm and RAOB's results





Figure 3. Histogram of two results of Grody's and Wang's TWV algoritms

  

Figure 4. Upper left are sever Typhoon Bilis (0010), announced eye of typhoon is located on 22.2N, 122.2E. Upper right is structure of typhoon center. Lower left is weak typhoon Kaitak(0004), announced eye of typhoon is located on 18.7N, 120.7E