M.SC Project Report

Rainfall Estimation by Remote Sensing
Technique
RAIN FALL ESTIMATION BY REMOTE SENSING TECHNIQUE
________________________________________________________________________
ABSTRACT
Flood and drought monitoring, management of agricultural crops, forecasting hydroelectric

power, assessments of extreme precipitation events, and other water resources management
applications all require reliable precipitation inputs.
The study of present work is to estimate rainfall using Remote sensing technique. Here we
are using Tropical Rainfall Measuring Mission (TRMM) satellite. For near real-time rainfall
estimation based on a Precipitation Radar (PR) onboard the Tropical Rainfall Measuring
Mission (TRMM) satellite. The Precipitation Radar (PR), an active sensor, is the first space-
based precipitation radar. PR emits radar pulses toward Earth, which are then reflected by
precipitation particles back to the radar. By measuring the strength of the returned pulses, the
radar is able to estimate rainfall rates.
The frequency of TRMM PR is 13.8 GHz.The PR can achieve quantitative rainfall estimation
over land as well as ocean. At present study PR (2A25) Algorithm is used for rainfall
estimation. The objectives of 2A25 algorithm are to correct for the rain attenuation in
measured radar reflectivity and to estimate the instantaneous three-dimensional distribution
of rain from the TRMM Precipitation Radar (PR) data.
The data of TRMM-PR is in HDF format. The Hierarchical Data Format, or HDF, is a
multi object file format for sharing scientific data in a distributed environment. In this study
TRMM-PR data processed by using some freely downloadable software TSDIS Orbit Viewer
and C & FORTRAN program. Finally we get ASCII file as output data.
In present study normalized TRMM data was also validated with respect to IMD data. For
validation process, Solapur area is selected. The polynomial regression analysis resulted 46%
of confidence level (R= .46). The best fit was obtained with higher polynomial order. This
indicates that there is some data anomaly within both the data sets. Due to gap in IMD data
and measurement of very fine atmospheric condition by TRMM produced data anomaly.
2
________________________________________________________________________
ACKNOWLEDGEMENT
My first and foremost sincere gratitude is to Thy Lord for showering thy blessings
upon me throughout the course of my project
At the outset I express my humble sense of gratitude and indebtedness to Dr.D.C
Gupta, Head of the Department of applied Geology.
I feel it as a special responsibility to convey my special thanks to my thesis supervisor

Dr. Siva Subramaniam & Dr. Brijendra Pateriya for their kind permission, to carry out
this project work at RMSI, Noida. I am also thankful to them for their constant moral
encouragement, tremendous amount of inspiration and affection during my Dissertation.
I express my deep sense of gratitude and regards to Sri Rajesh. S. Paul & Sri
Subhashbabu Ainampudi for their affectionate guidance, continuous encouragement and
inspiration given throughout my project. I am not exaggerating if I say there is no word in my
dictionary to express my regards towards them.
I wish to express my heartfelt thanks to Mr. Rajesh Kumar, Research Assistant,

M.N. Saha Center of Space Studies, University of Allahabad for his affection and
encouragement during this thesis work.
I feel it is special privilege to egress my heartfelt thanks to Ms. Niva Srivastava, Mr.
Arun Kumar Balla and Mr. Manohar Mishra for their lots of support.
I wish to express my deep sense of gratitude to my parents and special thanks to my

family members for their encouragement and support in my educational career.
I do not have words to express my thanks to my classmates, friends and well- wishers
who helped me and made my stay at RMSI Noida a memorable and pleasant experience.
SUNITA KUMARI
3
________________________________________________________________________
LIST OF CONTENTS
1 INTRODUCTION ............................................................................................................ 7
1.1 Introduction of Remote sensing ...............................................................................................................7
1.2 Types of Remote Sensing.........................................................................................................................7
1.3 Bands Used in Remote Sensing ...............................................................................................................7
1.4 Spectral Characteristics vis-à-vis Different Systems................................................................................8
1.5 Energy Interactions, Spectral Reflectance and Colour Readability in Satellite Imagery .......................10
1.5.1 Energy interactions ......................................................................................... 11
1.5.2 Spectral reflectance and color readability....................................................... 11
1.6 Planck radiation law ...............................................................................................................................11
1.7 Objectives...............................................................................................................................................12
Rainfall estimation by remote sensing technique using TRMM satellite data. ...... 12
Validation of TRMM data with respect to IMD ..................................................... 12
2 LITERATURE REVIEW ............................................................................................... 13
2.1 General ...................................................................................................................................................13
2.2 Rain Gauge- Measurement.....................................................................................................................13
2.3 WSR-88D Precipitation Products...........................................................................................................13
2.4 Radar Application ..................................................................................................................................14
2.5 The Operational GOES Infrared Rainfall Estimation.............................................................................14
2.6 Tropical Rainfall Measuring Mission (TRMM).....................................................................................15
2.6.1 Advantages...................................................................................................... 15
2.6.2 Limitation........................................................................................................ 16
3 METHODOLOGY AND SOURCE OF DATA............................................................. 17
3.1 General Introduction About TRMM ......................................................................................................17
3.2 Scope......................................................................................................................................................17
3.3 TRMM Instruments:- .............................................................................................................................21
3.3.1 TRMM Microwave Imager (TMI).................................................................. 21
3.3.2 Precipitation Radar (PR) ................................................................................. 21
3.3.3 Visible Infrared Scanner (VIRS) .................................................................... 21
3.3.4 Cloud and Earth’s Radiant Energy System (CERES)..................................... 21
3.3.5 Lightning Imaging Sensor (LIS) ..................................................................... 22
3.4 TRMM precipitation radar system description.......................................................................................22
3.4.1 TRMM-PR Algorithms................................................................................... 26
3.4.2 Algorithm Overview ....................................................................................... 30
3.4.3 More Detailed Description of the Algorithm.................................................. 31
3.5 TRMM Data format ...............................................................................................................................34
3.5.1 Introduction to HDF........................................................................................ 34
3.5.2 Primary HDF Platforms .................................................................................. 36
3.6 Methodology ..........................................................................................................................................36
3.6.1 Software .......................................................................................................... 38
3.6.2 GrADS1.9 b4 Linux software ......................................................................... 51
3.6.3 HDF Library.................................................................................................... 53
4 RESULT AND VALIDATION ...................................................................................... 69
4.1 General Discussion.................................................................................................................................69
4.2 Validation...............................................................................................................................................70
4.2.1 Conclusion ...................................................................................................... 72
4.2.2 Future scope .................................................................................................... 72
5 CONCLUSION............................................................................................................... 73
4
________________________________________________________________________
LIST OF FIGURE
Figure 1.1 Various types of platforms in Remote Sensing .................................................. 9

Figure 1.2 The electromagnetic spectrum............................................................................... 10
Figure 1.3 Basic interactions between electromagnetic energy and an earth surface feature. 11
Figure 3.1 TRMM orbit path ............................................................................................... 20
Figure 3.2 TRMM Orbit paths ............................................................................................ 20
Figure 3.3 Shown TRMM sensor scanning system ............................................................ 22
Figure 3.4 Measurement Concept Of Precipitation Radar ............................................... 23
Figure 3.5 T RMM Precipitation Radar Algorithm Flow ....................................................... 30
Figure 3.6 HDF Data Structure ............................................................................................... 35
Figure 3.7 Primary HDF Platform .......................................................................................... 36
Figure 3.8 Flow Process Diagram........................................................................................... 37
Figure 3.9 The Main Window of TSDIS orbit viewer ............................................................ 39
Figure 3.10 The Table Viewer ................................................................................................ 40
Figure 3.11 Meta data ............................................................................................................. 41
Figure 3.12 Fixed cross section .............................................................................................. 42
Figure 3.13 Variable Cross Section ........................................................................................ 43
Figure 3.14 Select 3D Locations............................................................................................. 44
Figure 3.15 3D control window (top), and 3D Image window (bottom)................................ 45
Figure 3.16 Definition of the sliders in the 3D Control window ............................................ 46
Figure 3.17 Generic View Control Window........................................................................... 47
Figure 3.18 The Generic View Text window ........................................................................ 48
Figure 3.19 Histogram window .............................................................................................. 48
Figure 3.20 A portion of a file summary ................................................................................ 49
Figure 3.21TRMM Mission index window ............................................................................ 50
Figure 3.22 Daily Time Series Window ................................................................................. 51
Figure 3.23 Using the TRMM Mission Index to Find Data of Interest............................ 51
Figure 4.1 show the relation between TRMM and IMD mean monthly rainfall data ............ 72
5
________________________________________________________________________
LIST OF TABLES
Table 1.1 subdivision of optical wave length region ................................................................ 9

Table 3.1 Main Characteristics of the TRMM Satellite.......................................................... 19
Table 3.2 Major Parameters of TRMM PR ............................................................................ 26
Table 3.3 Major changes of PR algorithm after the last version (version 6.0) ....................... 29
Table 4.1 Surface Rainfall ...................................................................................................... 70
Table 4.2 shows mean monthly rainfall data of TRMM and IMD ........................................ 71
6
________________________________________________________________________
CHAPTER 1
1 INTRODUCTION
1 .1 Introduction of Remote sensing
Remote Sensing is the science and art of acquiring information (spectral, spatial, and
temporal) about material objects, area, or phenomenon, without coming into physical contact
with the objects, or area, or phenomenon under investigation. Without direct contact, some
means of transferring information through space must be utilized. In remote sensing,
information transfer is accomplished by use of electromagnetic radiation (EMR). EMR is a
form of energy that reveals its presence by the observable effects it produces when it strikes
the matter.
1 .2 Types of Remote Sensing
Remote sensing can be classified on the basis of energy resources and the wavelength regions
as mentioned below
In respect to the type of Energy Resources
 Passive Remote Sensing: In this branch of remote sensing sensors detect the reflected or
emitted electro-magnetic radiation from natural sources, like Sun. Optical remote sensing
comes under this category.
 Active Remote Sensing: In such type of remote sensing sensors detect reflected
responses from the objects that are irradiated from artificially generated energy sources,
such as radar. Microwave comes under this type of remote sensing.
In respect to Wavelength Regions
Remote Sensing is classified into three types in respect to the wavelength regions
 Visible and Reflective Infrared Remote Sensing

 Thermal Infrared Remote Sensing
 Microwave Remote Sensing
1 .3 Bands Used in Remote Sensing
Emission of EMR from gases is due to atoms and molecules in the gas. Atoms consist of a
positively charged nucleus surrounded by orbiting electrons, which have discrete energy
7
________________________________________________________________________
states. Transition of electrons from one energy state to the other leads to emission of
radiation at discrete wavelengths. The resulting spectrum is called line spectrum. The
wavelengths, which are emitted by atoms/molecules, are also the ones, which are absorbed
by them. Emission from solids and liquids occurs when they are heated. Due to this heat a
continuous spectrum generated. This is called thermal emission and it is an important source
of EMR from the viewpoint of remote sensing. EMR, which is reflected or emitted from an
object, is the usual source of Remote Sensing data. However, any medium, such as gravity or
magnetic fields, can be used in remote sensing. Remote sensing technology makes use of
various types of platforms with sensors on board as shown in Fig. 1.1.
The wide range Electro-Magnetic Spectrum (EMS) from a very short wave "Gamma Ray" to
a very long 'Radio Wave'. Wavelength regions of electro-magnetic radiation have different
names ranging from Gamma ray, X-ray, Ultraviolet (UV), Visible light, Infrared (IR) to
Radio Wave, in order from the shorter wavelengths. The optical wavelength region, an
important region for remote sensing applications, is further subdivided as shown in Table 1.1
the electromagnetic spectrum shown in Fig 1.2.
Microwave region (1mm to 1m) is another portion of EM spectrum that is frequently used to
gather valuable remote sensing information.
1 .4 Spectral Characteristics vis-à-vis Different Systems
The sunlight transmission through the atmosphere is effected by absorption and scattering of
atmospheric molecules and aerosols. This reduction of the sunlight's intensity is called
extinction.
8
________________________________________________________________________
Figure 1.1 Various types of platforms in Remote Sensing
Name Wavelength (mm)

Opticalwavelength
Reflective portion 0.30-15.0
(i)Visible 0.38-3.00
(ii)NearIR 0.38-0.72
(iii) 0.72-1.30
Middle IR 1.30-3.00
Far IR (Thermal, 7.00-15.0
Emissive)
Table 1.1 subdivision of optical wave length region
9
________________________________________________________________________
Figure 1.2 The electromagnetic spectrum
One cannot select the sensors to be used in any given remote-sensing task arbitrarily; one
must instead consider the available spectral sensitivity of the sensors:
1. The presence or absence of atmospheric windows in the spectral range(s) in which one
wishes to sense, the source, magnitude, and spectral composition of the energy available
in these ranges.
2. The choice of spectral range of the sensor must be based on the manner in which the
energy interacts with the features under investigation.
1 .5 Energy Interactions, Spectral Reflectance and Colour Readability in Satellite Imagery
All matter is composed of atoms and molecules with particular compositions. Therefore,
matter will emit or absorb electro-magnetic radiation on a particular wavelength with respect
to the inner state. All matter reflects, absorbs, penetrates and emits Electro-magnetic
radiation in a unique way. For example, the reason why a leaf looks green is that the
chlorophyll absorbs blue and red spectra and reflects the green. The unique characteristics of
matter are called spectral characteristics.
10
________________________________________________________________________
1 .5 .1 Energy interactions
When electro-magnetic energy is incident on any given earth surface feature, three
fundamental energy interactions with the feature are possible (Figure. 1.3)
Figure 1.3 Basic interactions between electromagnetic energy and an earth surface
feature
1 .5 .2 Spectral reflectance and color readability
Two points about the above given relationship (expressed in the form of equation) should be
noted :
 The proportions of energy reflected, absorbed, and transmitted will vary for different
earth features, depending upon their material type and conditions. These differences
permit us to distinguish different features on an image.
 The wavelength dependency means that, even within a given feature type, the
proportion of reflected, absorbed, and transmitted energy will vary at different
wavelengths.
Thus, two features may be distinguishable in one spectral range and be very different on
another wavelength brand. Within the visible portion of the spectrum, these spectral
variations result in the visual effect called COLOUR. For example we call blue objects 'blue'
when they reflect highly in the 'green' spectral region, and so on. Thus the eye uses spectral
variations in the magnitude of reflected energy to discriminate between various objects.
1 .6 Planck radiation law
The primary law governing blackbody radiation is the Planck Radiation Law, which governs
the intensity of radiation emitted by unit surface area into a fixed direction (solid angle) from
11
________________________________________________________________________
the blackbody as a function of wavelength for a fixed temperature. The Planck Law can be
expressed through the following equation.
h
E hc / kT
e 1
h = 6.625X 10-27 erg-sec (Planck Constant)

k = 1.38X10-16 erg-sec (Boltzman Constant)
c = 3X1010 cm/sec (Speed of Light)
Planck's law is strictly valid only for ideal blackbodies, which by definition have 100 %
absorption and maximum emitting intensity (i.e. existence). Real objects can be handled by
introducing emissivity, a factor less than unity, equal to the ratio of the emitting intensity of
the object and that of a corresponding blackbody having the same temperature as the object.
In the general case the emissivity is wavelength dependent. Objects with a less than unity and
almost constant emissivity are named grey-bodies.
1 .7 Objectives
1. Rainfall estimation by remote sensing technique using TRMM satellite data.
2. Validation of TRMM data with respect to IMD
12
________________________________________________________________________
CHAPTER 2
2 LITERATURE REVIEW
2 .1 General
For the present study, an exhaustive review of the available literature on the following topics
is done.
2 .2 Rain Gauge- Measurement
Rain gauge is one of the oldest methods for estimating the rainfall. The rain gauge long
service and widespread use has made it the standard for measuring surface rainfall (Hunter,
1996). However, there is error in gauge measurement especially due to turbulence and wind
flow losses (Yang et al., 1998). Traditionally, taking rainfall measurements has been a simple
matter of setting out rain gauges. Around 350 BC, the Indian author, Kautilya, described the
first known rain gauge in his manuscript, Arthasastra. Described as a 20-inch diameter bowl,
measurements were taken regularly as a way of estimating the annual crop to be sown. The
Chinese reported the first known use of rain gauges in the year 1247 AD during the Southern
SongDynasty.
If you want to measure the amount of rain where you live, you can simply set out a rain
gauge and collect the rain that is falling near your house. Simple and inexpensive rain gauges
will let you measure rain amounts to the nearest 1/100th of an inch. Researchers that need to
measure the rainfall for an entire area will set out a network of rain gauges, collect the data
and map out the results. This tedious task is adequate for specialized projects, but does not
work well when trying to map out rainfall for the entire world. The advent of weather radar
changed this problem/situation considerably. Today, single weather radar can map out
rainfall in a circle that may have a diameter of 250 miles (400 kilometers). Wilson and
Brandes (1979) determined that wind/turbulence errors are usually around 5%, but with high
winds in thunderstorm outflows can be as high as 40%. The primary problem is that the
gauge measurement is for a point, whereas, the radar yields a rainfall estimate for a much
larger area. Distances between National Weather Service gauges can often exceed 60
miles. Estimating or interpolating precipitation events from these scattered gauge
observations can misrepresent the event, especially convective showers and even large-scale
stratiform precipitation events. The WSR-88D radar certainly has an advantage in providing
a surface precipitation estimate because of the high spatial and temporal resolution
2 .3 WSR-88D Precipitation Products
13
________________________________________________________________________
The WSR-88D Doppler radar is an S-band transmitter that records the reflectivity, radial
velocity and spectrum width of reflected signals. This is the WSR-88D base data produced
by the Radar Data Acquisition (RDA) unit. Data at this stage is referred to archive level II.
The next stage in the processing stream is the Radar Product Generator (RPG), which applies
computer algorithms to produce meteorological and hydrological products. Data from this
stage is referred to archive level III. A more complete description of these data products and
processing may be found in Crum and Alberty (1993). The unprocessed reflectivity obtained
from level II data is converted to rainfall rates using a relationship between reflectivity and
rainfall rate called the Z-R relationship. The base reflectivity is collected in a radial
coordinate system during 360-degree azimuthal scans in ranges of 1 km and sectors of one
degree. Each scan is incremented 0.5 degree in elevation to form a volume scan. Depending
on the number of tilt angles used, a volume scan takes from 5 to 10 minutes to complete.
During the RPG processing, the data is aggregated in time and space depending on the
product generated. The precipitation products generally available to the public are at an
aggregated resolutions and time intervals. The NWS River Forecast Centers (RFC) forecast
river stage using a mosaic of radars aggregated to 4 km and 6 hour spatial/temporal
resolution. This product is known as stage III that is used by the RFC to forecast river stage.
2 .4 Radar Application
Accurate observations of precipitation (e.g., rain and snow) are very important to
hydrological applications. Historically surface rainfall has been measured using rain gauges
deployed in networks of various gauge density. With improved remote sensing capabilities,
particularly radar, these point measures of rainfall have increasingly been augmented by
area-covering, remotely sensed rain estimates. Radar enables three-dimensional observation
of precipitation with excellent aerial coverage and high resolution in space and time. The
radar reflectivity factor Z is physically related to rainfall rate R. For this study we make use
of data collected in the Goodwin Creek research watershed. Since early 1996, we have
collected approximately eighty rainfall events with accumulations in excess of 10 mm each.
For most of these storms archive level II data of the NEXRAD WSR-88D located at
Memphis (Tennessee) have been obtained. These data, in combination with the rain gauges
of the Goodwin Creek watershed and the dichromate, is used to investigate the problems
involved in the rainfall measurement and develop strategies on how to obtain accurate
surface rainfall estimates. This study highlights the importance of carefully designed field
observations and the need for much more rigorous analysis of precipitation as it reaches the
ground than we have done in the past.
2 .5 The Operational GOES Infrared Rainfall Estimation
This paper presents a description, sensitivity analyses ample results, validation, and the
recent progress done the development of a new satellite rainfall estimation technique in the
National Environmental Satellite Data and Information Service (NESDIS) at the National
Oceanic and Atmospheric administration (NOAA). The technique, called the auto-estimator,
runs in real time for applications to flash flood forecasting, numerical modeling, and
operational hydrology. The auto-estimator uses the Geoestationary Operational
Environmental Satellite-8 and -9 in the infrared (IR) 10.7-mm band to compute real-time
precipitation amounts based on a power-law regression algorithm. This regression is derived
14
________________________________________________________________________
from a statistical analysis between surface radar–derived instantaneous rainfall estimates and
satellite-derived IR cloud-top temperatures collocated in time and space. The rainfall rate
estimates are adjusted for different moisture regimes using the most recent fields of
perceptible water and relative humidity generated by the National Centers for Environmental
Prediction Eat Model. In addition, a mask is computed to restrict rain to regions satisfying
two criteria: (a) the growth rate of the cloud as a function of the temperature change of the
cloud tops in two consecutive IR mages must be positive, and (b) the spatial gradients of the
cloud-top temperature field must show distinct and isolated cold cores in the cloud-top
surface. Both the growth rate and the gradient corrections are useful for locating heavy
precipitation cores. The auto-estimator has been used experimentally for almost 3 yr to
provide real-time instantaneous rainfall rate estimates, average hourly estimates, and 3-, 6-,
and 24-h accumulations over the conterminous 48 United States and nearby ocean areas. The
NOAA/NESDIS Satellite Analyses Branch (SAB) has examined the accuracy of the rainfall
estimates daily for a variety of storm systems. They have determined that the algorithm
produces useful 1–6-h estimates for flash flood monitoring but exaggerates the area of
precipitation causing overestimation of 24-h rainfall total associated with slow-moving, cold-
topped musicale convective systems. The SAB analyses have also shown a tendency for
underestimation events should be considered with caution. The authors validate the hourly
rainfall rates of the auto-estimator using gauge-adjusted radar precipitation products (with
radar bias removed) in three distinct cases. Results show that the auto-estimator has modest
skill at 1-h time resolution for a spatial resolution of 12 km. Results improve with larger grid
sizes (48 by 48 km or larger).
2 .6 Tropical Rainfall Measuring Mission (TRMM)
TRMM satellite is most advance technique for rainfall measurement. Tropical Rainfall
Measuring Mission (TRMM) satellite project, established the first Active Microwave
Precipitation Radar (PR) onboard TRMM as the first space weather surveillance radar
station. The excellent conceptual designs give TRMM the merit to observe vertical rain
profile from the top of cloud propagate until the surface effectively. TRMM-PR gives direct
rainfall. Result of this sensor is very good as compare to other using technique. People have
been measuring rainfall amounts for more than 2000 years, but we still don't know how much
rain falls in remote areas of the globe. For the first time, we will be able to directly measure
global rainfall rates with the Tropical Rainfall Measuring Mission (TRMM). Floods and
droughts have been making history since the beginning of civilization. Rainfall, however,
also is associated with the less catastrophic uncertainties that affect us all. Farmers know
that? you have to make hay when the sun shines.? If farmers want to harvest their hay, it
must be dry, and they sometimes work all day and all night to bring in their crops if rain is in
the forecast. In the larger economic picture, not only the farmers, but also economists and
financial traders monitor rainfall to see if there will be healthy crops. If production is
abundant, prices will probably drop and investors can plan their investments accordingly.
2 .6 .1 Advantages
The Precipitation Radar will be the first space borne instrument designed to provide three-
dimensional maps of storm structure. The measurements should yield invaluable information
on the intensity and distribution of the rain, on the rain type, on the storm depth and on the
15
________________________________________________________________________
height at which the snow melts into rain. The estimates of the heat released into the
atmosphere at different heights based on these measurements can be used to improve models
of the global atmospheric circulation.
The Precipitation Radar has a horizontal resolution at the ground of about 2.5 miles (four
kilometers) and a swath width of 137 miles (220 kilometers). One of its most important
features is its ability to provide vertical profiles of the rain and snow from the surface up to a
height of about 12 miles (20 kilometers). The Precipitation Radar is able to detect fairly light
rain rates down to about .027 inches (0.7 millimeters) per hour. At intense rain rates, where
the attenuation effects can be strong, new methods of data processing have been developed
that help correct for this effect. The Precipitation Radar is able to separate out rain echoes for
vertical sample sizes of about 820 feet (250 meters) when looking straight down. It will carry
out all these measurements while using only 224 watts of electric power—the power of just a
few household light bulbs. The Precipitation Radar was built by the National Space
Development Agency (JAXA) of Japan as part of its contribution to the joint US/Japan
Tropical Rainfall Measuring Mission (TRMM).
Although weather radar technology enables us to map rainfall over land, it is not likely to be
found in less developed countries or in sparsely populated areas of the world due to its
expense, nor is it capable of measuring rainfall over the oceans. Weather radars do exist on
some islands and ships, but the coverage is much less than adequate. To address this
problem, the United States Department of Defense has established a constellation of polar
orbiting satellites that carry Special Sensor Microwave/Imager (SSM/I) instruments. These
sensors are called Passive detectors, as they are able to simply receive microwave energy that
is emitted naturally from the Earth below. Variations in the amount of energy received at
different frequencies can be interpreted in terms of the rate at which the rain is falling near
the ground while the satellite is overhead
2 .6 .2 Limitation
There is one limitation in TRMM satellite mission. Per day it rotation is 15 times, and it
cover entire world. Somewhere suppose heavy rainfall is occurred but that time if it not
passes from that area we can't estimate the rainfall of that area.
16
________________________________________________________________________
CHAPTER 3
3 METHODOLOGY AND SOURCE OF DATA
3 .1 General Introduction About TRMM
TRMM, Tropical Rainfall Measuring Mission, was launched by the H-II rocket from
Tanegashima Space Center of NASDA, at AM 6:27 JST, November 28, 1997. This satellite
has been developed as a joint project between Japan and US, which is the first space mission
dedicated to measuring rainfall. TRMM will mainly observe rain structure, rate and
distribution in tropical and subtropical region, the data is expected to play an important roll
for understanding mechanisms of global climate change and monitoring environmental
variation.
3 .2 Sc o pe
Decision-makers and researchers require accurate spatial and temporal precipitation

estimates. High-resolution flood and drought monitoring, management of agricultural crops,
forecasting hydroelectric power, assessments of extreme precipitation events, and other water
resources management applications all require reliable precipitation inputs. For a century,
the rain gauge has been the standard for measuring surface rainfall and is often assumed to be
"ground truth" because of its long service and widespread use (Hunter, 1996). However,
gauges produce point measurements that are assumed to represent a lager area.
The United States meteorological network consists of about 8,000 daily rainfall-recording
stations (Groisman and Legates, 1994). Distances between stations often exceed 60 miles
(Real-Time Calibration of Radar Precipitation Estimates, 1998) providing inadequate spatial
sampling. Gauge measurement errors produced by wind/turbulent losses (Yang et al. 1998;
Groisman and Legates, 1994), gauge wetting, into and out of the gauge (Brandes and Wilson,
1998), and evaporation (Austin, 1987) compound estimation.
Rainfall is one of the most important weather and climate variables that determine whether
mankind survives, thrives, or perishes. Water is so ubiquitous on planet Earth that we often
take it for granted. Too much water results in devastating floods, and the famine caused by
too little water (drought) is responsible for more human deaths than all other natural disasters
combined. Water comprises more than 75 percent of our bodies and as much as 95 percent of
some of the foods we eat. Water is essential to life, as it nourishes our cells and removes the
waste they generate. Water determines whether plants produce food, or whether they wither
from drought or rot from dampness. Water is essential to our homes and factories, to our
production of food, fiber, and manufactured goods, and to just about everything else we
produce and consume. Although water covers more than 70 percent of the Earth’s surface,
only about 3 percent is fresh water. And about 75 percent of that is inaccessible because it is
locked up in glaciers and ice caps. Another important aspect of rainfall, or any other
17
________________________________________________________________________
precipitation, is its role in redistributing the energy the Earth receives from the Sun.
Evaporation of water from the Earth’s surface, condensation of water vapor into cloud
droplets or ice particles, snow, precipitation, runoff of the precipitation, and melting of snow
and ice constitute what is known as the water cycle, or the hydrological cycle. Evaporation,
the process of changing water from liquid to gas form, absorbs 540 calories of energy per
gram of water, while simply raising the temperature of a gram of water by one degree celsius
without changing its phase, requires only one calorie of energy. Thus, much of the Sun’s
energy that reaches the Earth’s surface is used to evaporate water instead of raising the
temperature of the surface. The resulting water vapor is carried upward by the atmosphere
until it reaches a level where it is cooled to its condensation temperature. Then the water
vapor releases the energy (540 calories per gram) it absorbed during the evaporation process.
This “latent heat” release can occur thousands of kilometers from where the latent heat was
originally absorbed. Water plays an additional critical role in weather and climate: water
vapor, it turns out, is the most abundant and most important greenhouse gas! Greenhouse
gases trap some of the energy given off by the Earth’s surface in the atmosphere. Therefore,
the distribution and quantity of water vapor in the atmosphere are important in determining
how well the Earth can emit the energy it absorbs from the Sun back into space. Unless the
Earth loses as much energy as it receives, it will warm up. If the Earth loses more energy
than it receives, it will cool down. The distribution of water vapor in the atmosphere also
affects cloudiness; and clouds play an important role in determining how much solar energy
reaches the Earth’s surface, as well as how much heat can escape to space Perhaps it is now
obvious that water, in all its forms, plays a critical role in determining what we call weather
and climate. Our understanding of the complicated interactions involving water is insufficient
to permit us to forecast, with much skill, weather beyond several days and climate beyond a
few months. Because the occurrence of precipitation is highly variable in both time and
space, and almost three-fourths of the Earth’s surface has no rain gauges because the oceans
cover it, we have never been able to adequately observe the global distribution of rain.
Measurements from rain gauges on islands and satellite images of clouds have led to
estimates of global precipitation. TRMM is the first satellite to measure precipitation with the
accuracy available from radar in combination with other remote sensors and it represents a
breakthrough in our ability to monitor precipitation on a global scale.
18
________________________________________________________________________
Launch weight Approx. 3.62 ton
Launcher H-II Rocket

Launch date (JST) November 28, 1997
Altitude Approx. 350 km
Inclination Approx. 35 degrees
Shape At lift-off: 5.1 m (length), 3.7 m

(diameter)
In orbit: 5.1 m (length), 14.6 m
(in paddle direction)
Total: 3620 kg [3524 kg]
Fuel: 890 kg
Dry weight: 2730 kg [2634 kg]
Weight Fuel: 890 kg

Dry weight: 2730 kg [2634 kg]
Power Approx. 1100 W [Ave. 850 W]
Attitude control Zero momentum three-axis

stabilized
Data transmission
Via TDRS
32Kbps (Real Time), 2Mbps
(Play Back)
Design life 3 years and 2 months

Mission Precipitation Radar (PR)
instrument TRMM Microwave Imager
(TMl)
Visible and Infrared Scanner
(VIRS)
Clouds and the Earth’s Radiant
Energy System (CERES)
Lightning Imaging Sensor (LIS)
Table 3.1 Main Characteristics of the TRMM Satellite
19
________________________________________________________________________
Figure 3.1 TRMM orbit path
Figure 3.2 TRMM Orbit paths
20
________________________________________________________________________
3 .3 TRMM Instruments:-
3 .3 .1 TRMM Microwave Imager (TMI)
The TRMM Microwave Imager (TMI) is a passive microwave sensor that detects and images
microwave radiation emitted by water droplets, ice particles, and the Earth’s surface. TMI
detects radiation at five different frequencies, which helps to distinguish between rainfall,
bodies of water, and land. Data obtained from this instrument is used to quantify the water
vapor, cloud water, and rainfall intensity in the atmosphere.
3 .3 .2 Precipitation Radar (PR)
The Precipitation Radar (PR), an active sensor, is the first space-based precipitation radar. PR
emits radar pulses toward Earth, which are then reflected by precipitation particles back to
the radar. By measuring the strength of the returned pulses, the radar is able to estimate
rainfall rates. Among the three main instruments on TRMM, PR is the most innovative.
Other instruments similar to TMI and the visible and Infrared Scanner (VIRS) have operated
in space before, but PR is the first radar launched into space for the purpose of measuring
rainfall. Data obtained from this instrument:
 Provides three-dimensional storm structures;
Helps to determine the intensity and three-dimensional distribution of rainfall over
land and Water, which can be used to infer the three-dimensional distribution of
latent heat in the atmosphere.
 Provides information on storm depth; and
 Provides information on the height at which falling snow or ice particles melt into
rain.
3 .3 .3 Visible Infrared Scanner (VIRS)
The Visible and Infrared Scanner (VIRS) measure radiance in five wavelength bands (from
visible to Infrared) emitted by clouds, water vapor and the Earth’s surface. The intensity of
radiation from a cloud corresponds with the brightness or temperature of the cloud, which
in turn indicates the height of the cloud. Brighter (colder) clouds are higher in altitude, and
darker (warmer) clouds are lower. In general, higher clouds are associated with heavier rain.
By comparing VIRS observations with rainfall estimates from TMI and PR, scientists are
able to better understand the relationship between cloud height and rainfall rate, and can
apply this knowledge to radiation measurements made by other weather satellites.
3 .3 .4 Cloud and Earth’s Radiant Energy System (CERES)
The Clouds and Earth’s Radiant Energy System (CERES) measure the amount of energy
rising from the Earth’s surface, atmosphere, and clouds. Clouds can have both a warning and
cooling effect on the Earth, trapping energy emitted by the Earth’s surface while blocking
21
________________________________________________________________________
energy from the Sun. Similarly, water vapor also warms the Earth by trapping outgoing
radiation, but also condenses to form clouds that sometimes have a cooling effect. Data from
this instrument helps scientists learn more about how the Earth distributes the energy it
receives from the Sun, as well as the effects of clouds and water vapor on the overall
temperature and energy budget of the Earth. This information will help long-term climate
models make more accurate predictions.
3 .3 .5 Lightning Imaging Sensor (LIS)
The Lightning Imaging Sensor (LIS) is a powerful instrument that can detect and locate
cloud-to ground, cloud-to-cloud, and intra-cloud lightning. The information gained from this
instrument is used to classify class type and, together with other TRMM instrument, to
correlate lightning flash rate with storm properties, including rainfall rate. It’s also expected
that the information provided from LIS will leads to future advances in lighting detection and
forecasting.
Figure 3.3 Shown TRMM sensor scanning system
3 .4 TRMM precipitation radar system description
The TRMM precipitation radar (PR) is the first space borne rain radar and the only
instrument on TRMM that can directly observe vertical distributions of rain. The frequency
of TRMM PR is 13.8 GHz. The PR can achieve quantitative rainfall estimation over land as
well as ocean. The PR can also provide rain height Information, which is useful for the
radiometer-based rain rate retrieval algorithms. The footprint size of PR is small enough to
allow for the study of inhomogeneous rainfall effects upon the comparatively coarse
footprints of the low frequency microwave radiometer channels. During the normal
observation mode, PR antenna beam scans in the cross-track direction over ±17° to results
22
________________________________________________________________________
220 km swath width from end to end. The antenna beam width of the PR is 0.71° and there
are 49 observation angle bins within the scanning angle of ±17°. The horizontal resolution
(footprint size) is 4.3 km at nadir and about 5 km at the scan edge when TRMM takes the
nominal altitude of 350 km. The range resolution of TRMM PR is 250 m, which is equal to
the vertical resolution at nadir. The radar echo sampling is performed over the range gates
between the sea surface and the altitude of 15 km for each observation angle bin. For nadir
incidence, the “mirror image” is also collected up to the altitude of 5 km. In addition,
“oversample ” echo data are partially collected for surface return echoes (for scan angle
within ±9.94°) and for rain echoes (for scan angles within ±3.55° up to the height of 7.5 km).
The minimum detectable Z (corresponding to the noise-equivalent received power) improved
from 23.3 dBZ (based upon the specifications requirement) to 20.8 dBZ as determined from
the pre-launch ground test and from the orbit test. This is mainly due to the increased
transmit power and the decrease of the receiver noise figure.
Actually the rain echo power is measured from the subtraction of the system noise power
from the total receiver power (rain echo power + system noise power). The accuracy of rain
echo power can be characterized by the effective signal-to-noise ratio (S/N), which is the
ratio of mean to standard deviation of rain echo power. By considering these facts, the actual
minimum detectable Z can be considered to be about 16-18 dBZ after the detailed statistical
calculation. The effective signal-to-noise ratio (S/N) of 3 dB is obtained when Z -factor is 17
dBZ.
Figure 3.4 Measurement Concept Of Precipitation Radar
23
________________________________________________________________________
The main performances of the precipitation radar are as follows:
1) Frequency : 13.796 GHz (f1) and 13.802 GHz (f2) (2 frequencies,

frequency agility)
2) Transmission : within ± 2 x 10-5/3 years and 2 months frequency stability
3) Occupied bandwidth: within 14 MHz
4) Spurious : 50 dBc or less (at antenna subsystem input/output port)
5) Range resolution : 250 m or less (normal at 6 dB width of reception filter
output pulse)
6) Horizontal resolution: 4.34 ± 0.12 km or less at nadir with a physical altitude of
350 km (Horizontal resolution on the ground where the
normal is 6 dB width of the transmit/receive round-trip
antenna pattern.)
7) Minimum radar echo: 111 dBm Reception level (Smin) (The value at the interface
point with the antenna subsystem. It is a reception level
where S/N for each pulse becomes 0 dB. Antenna input
noise temperature is assumed to be 290 K.)
8) Minimum measurable: 0.5 mm/h (S/N per pulse = 0 dB at the peak of rain area.)
rainfall intensity.
9) Scan width : During observation: 215 km or more (Between centers of

the width footprint at the Mode surface when geographical
altitude is at 350 km.)
10) Scan angle interval and the number of scan angle bins within each scan
: During observation: Scan angle interval: 2 scan angle bins
(about 0.71 degree) mode number of scan angle bins : 49
(including nadir)
11) Scan cycle : During observation: 0.6 seconds or less mode
12) Antenna orientation fix: During the external calibration mode, it is possible to fix
the antenna beam direction at the scan angle specified
within the scan angle bins, in addition to antenna scan
provided in 9) to 11).
13) Observation range
During observation : It is possible to observe surface echo by antenna main lobes
from an altitude of 20 km1. Furthermore, at a scan angle of
0°, a mirror image of up to an altitude of 5 km is included.
14) System noise level : Measured within the range at which the radar echo can be
ignored.
15) Averaged individual sample number of a radar video signal: 64 or more
16) Dynamic range : Because both the surface echo level and the noise level are
measured simultaneously at nadir (0 at sea surface shall be
16 dB, and the antenna input noise temperature shall be 120
K), the linear section of the receiver integrated input/output
characteristic (a characteristic where receive subsystem
noise is assumed to be ignorable detector, and A/D
conversion ) to be ignorable and includes logarithm and
24
________________________________________________________________________
includes logarithm detector and A/D conversion ) to the

sinusoidal pulse input has a margin of 5 dB or more above
and below noise of 5 dB or more above and below noise
and surface echo levels provided in this section.
17) Linearity : Within ± 0.6 dB in the linear section of 16).
18) Range reference point determination precision: 10 % or less of range resolution
19) Measurement precision
a) Equivalence Z-factor and surface normalized radar cross-section: within ± 1dB3, 4
b) Radar receive power : within ± 0.9 dB4
c) Radar transmit power: within ± 0.4 dB
1. It measures range up to 23 km originally, but only the data by 20 km will be
used finally.
2. It is an assumed point where the distance from the precipitation radar is
taken to be 0; and it shall be range bin 1.
3. Excludes error caused by rainfall and atmosphere attenuation.
4. Excludes error caused by statistical variation for each pulse in the radar
reception level. This provision is applied to the signal level within the
dynamic range provided in 16
.20) Surface echo strength: Less or equal to 0.5 mm/h rainfall echo strength within the
by the antenna side observation range provided in 13).
Antenna side lobes shall be lobes and the cross -20 dB or
less than the noise level, and cross polarization grating
polarization grating lobes shall be -5 dB or less than the
noise level in the system noise lobes detection window
provided in 14). Antenna input noise temperature shall be
120 K.
21) Range side lobes for receive pulse: -25 dB or less
22) Reception filter loss : 1.5 dB or less
23) Antenna gain : 47.4 dB or more (At antenna subsystem input/output port)
24) Antenna beam half- : 0.71 ± 0.02 degrees (at nadir) width 0.74 ± 0.03 degrees (at
a scan angle of 17 degrees)
25) Antenna side lobes : peak value -27 dB or less integration value at the same
range -64 dB or less (0°40°)
26) Inclination angle of antenna beam: 4 ± 0.1 degrees to the direction of feeding point
27) Maximum antenna scan angle: ± 17 degrees or more in the cross track direction
28) Grating lobes : Not generated when the antenna is scanned within the scan
angle range provided in 27).
29) Cross polarization grating lobes: -15 dB
30) Transmit/receive beam orientation conformity: within ± 0.07°
31) Beam orientation precision: The error of the antenna beam orientation (uncertainty) to
the radar alignment reference is within ± 0.2°.
32) Range bin
a) Number of range bins : 400
b) Range bin intervals : 125 m ± 1 m
33) Scan angle bin
25
________________________________________________________________________
a) Number of scan angle : 103

b) Scan angle bin intervals: 0.355° ± 0.1°
Table 3.2 Major Parameters of TRMM PR
3 .4 .1 TRMM-PR Algorithms
An algorithm is a computer program that processes data, not the set of mathematical steps
that the program carries out. The TRMM PR standard algorithms are developed by the
TRMM science team. They are classified into Level 1 (1B21, 1C21), Level 2 (2A21, 2A23,
2A25) and Level 3 (3A25, 3A26). Level 1 and Level 2 products are data in the IFOV. Level
3 data give the monthly statistical values of rain parameters mainly in 5° x 5° grid boxes
required by the TRMM mission.,. The algorithm 1B21 produces engineering values of radar
received power (signal + noise) and noise levels. It decides whether there exists rain or not in
26
________________________________________________________________________
the IFOV. It also estimates the effective storm height from the minimum detectable power
value. Algorithm 1C21 gives the radar reflectivity factor, Z, including rain attenuation
effects. The algorithm 2A21 computes the spatial and temporal statistics of the surface
scattering coefficient σ0 over ocean or land when no rain is present in the IFOV. Then,
when it rains in the IFOV, it estimates the path attenuation of the surface scattering
coefficient σ0 by rain using no rain surface scattering coefficient σ0 as a reference
[Meneghini, 2000]. The algorithm 2A23 tests whether a bright band exists in rain echoes and
determines the bright band height when it exists [Awaka, 1997] .The rain type is classified
into the stratiform type, convective type and others by the 2A23. It also detects shallow
isolated rain whose height is below the melting level height (zero degree Celsius). The
algorithm 2A25 retrieves profiles of the radar reflectivity factor, Z, with rain attenuation
correction and rain rate for each radar beam by the combination of Hitschfeld-Bordan and
surface reference methods [Iguchi,2000]. As the 13.8 GHz frequency band selected for the
TRMM PR is fairly heavily attenuated by rain, the compensation of this rain attenuation
becomes the major subject in the rain retrieval algorithms. Algorithm 3A25 gives the space-
time averages of accumulations of 1C21, 2A21, 2A23 and 2A25 products. The most
important output products are monthly averaged rain rates over 0.5° x 0.5° and 5° x 5° grid
boxes. It also outputs the monthly averaged bright band height over 0.5° x 0.5° and 5° x
5°grid boxes. Algorithm 3A26 gives monthly averaged rain rates over the 5° x 5° grid boxes.
PRODUCT NO NAME CONTACT PRODUCTS ALGORITHM

PERSON DESCRIPTION
1B21 PR calibration JAXA/EOC Total received Conversion of the

Rain/No Rain (Japan) power, Noise count value of
J. Awaka Level Clutter radar echoes and
(Japan) contamination noise level into
T. Iguchi flag. engineering value.
(Japan) Decision of
rain/no rain.
Determination of
effective storm
height from
minimum
detectable power
value. Rejection
of mainlobe and
sidelobe clutter.
1C21 PR JAXA/EOC Profiled Zm Conversion of the
Reflectivity (Japan) (radar power and noise
reflectivity value to radar
factors without reflectivity factors
rain attenuation Zm without rain
correction). attenuation
correction.
27
________________________________________________________________________
2A21 Surface R. Meneghini Path integrated Estimation of path

Scattering (USA) attenuation integrated
coefficient (PIA) of σ0 (in attenuation and its
case of rain) reliability using
and its the surface as a
reliability. Data reference target.
base of σ0 Spatial and
(ocean/ land, in temporal statistics
case of no rain) of surface σ0 and
classification of
σ0 into
land/ocean,
rain/no rain.
2A23 PR qualitative J. Awaka Detection of Whether a bright

(Japan) bright band, band exists in rain
Bright band echoes or not, and
height, strength, determination of
width, Rain bright band height
type when it exists.
classification, The rain type is
Detection of classified into
shallow isolated stratiform
rain. Output of type, convective
rain/no rain type or others.
flag, height of Shallow isolated
storm top. rain, the height of
which is below
the 0 deg., is
detected.
2A25 PR Profile T. Iguchi Range profiles The rainfall rate

(Japan) of estimate is given
attenuation- at each resolution
corrected cell. This
radar algorithm
reflectivity employs hybrid
factors, rainfall method of the
rate. The surface reference
estimated near method and
surface, and Hitschfeld-Bordan
surface rainfall method.
rate, and precipitation
average rainfall water content at 5
rate between altitudes, and
the two vertically
28
________________________________________________________________________
predefined integrated
altitude precipitation
(2,4 km). water content are
also
calculated.
3A25 Space time R. Meneghini Space-time Calculation of

average of (USA) averages of various statistics
radar product accumulations over a
of 1C21, month from the
2A21, 2A23, level 2 PR output
2A25. products. Four
types of statistics
are calculated.
1. probabilities of
occurrence,
2. means and
standard
deviations,
3. histograms,
4. correlation
coefficients.
3A26 Estimation of R. Meneghini Rain rate Estimated values

space time rain (USA) statistics over of the probability
rate statistics 5 degree x 5 distribution
degree 1 month function of the
space-time space-time
regions using a rain rate at 4
multiple levels, and the
threshold mean, standard
technique. deviation, and so
on.
Table 3.3 Major changes of PR algorithm after the last version (version 6.0)
29
________________________________________________________________________
Figure 3.5 T RMM Precipitation Radar Algorithm Flow
3.4.1.1 2A25
3.4.1.1.1 Objectives
The objectives of 2A25 are to correct for the rain attenuation in measured radar reflectivity
and to estimate the instantaneous three-dimensional distribution of rain from the TRMM
Precipitation Radar (PR) data. The estimated vertical profiles of attenuation-corrected radar
reflectivity factor and rainfall rate are given at each resolution cell of the PR. The estimated
rainfall rate at the actual surface height and the average rainfall rate between the two
predefined altitudes (2 and 4 km) are also calculated for each beam position.
3 .4 .2 Algorithm Overview
2A25 basically uses a hybrid of the Hitschfeld-Bordan method and the surface reference
method to estimate the vertical profile of attenuation-corrected effective radar reflectivity
factor (Ze). (The hybrid method is described in Iguchi and Meneghini (1994).) The vertical
rain profile is then calculated from the estimated Ze profile by using an appropriate Ze-R
relationship. One major difference from the method described in the above reference is that
in order to deal with the uncertainties in measurements of the scattering cross section of
surface as well as the rain echoes, a probabilistic method is used. Since the strong surface
30
________________________________________________________________________
echo hides radar rain echoes from near the surface, the rain estimate at the lowest point in the
clutter-free region is given as the near-surface rainfall rate for each angle bin.
3 .4 .3 More Detailed Description of the Algorithm
The major input data to 2A25 are the measured radar reflectivity factor Zm, the apparent
decrease of the surface cross section ( .σ 0), its reliability, the rain type and miscellaneous
height information. The algorithm first defines the region for processing: It processes only
the data between the rain top and the lowest height above the surface that is free from the
surface clutter. (The current algorithm does not use any data below the surface, i.e., the
mirror image.)
The bright-band height and climatological freezing height are used to define the regions of
liquid (water), solid (ice), and mixed phase of precipitating particles. The initial values of the
coefficients in the k-Ze and Ze-R relationships at different altitudes are accordingly defined.
The attenuation correction is, in principle, based on the surface reference method. This
method assumes that the decrease in the apparent surface cross section is caused by the
propagation loss in rain. The coefficient α in the k-Ze relationship k= αZe β is adjusted in such
a way that the path-integrated attenuation (PIA) estimated from the measured Zm-profile will
match the reduction of the apparent surface cross section. The attenuation correction of Ze is
carried out by the Hitschfeld-Bordan method with the modified α. Since α is adjusted, we
call this type of surface reference method the α-adjustment method. The α-adjustment
method assumes that the discrepancy between the PIA estimates from .σ 0and that from the
measured Zm-profile can be attributed to the deviation of the initial α value from the true
values, which may vary depending on the raindrop size distribution and other conditions. It
assumes that the radar is properly calibrated and that the measured Zm has no error.
The surface reference method generally works rather well as long as the apparent decrease in
surface cross section .σ 0 is much larger than the fluctuations of the true surface cross section.
When the decrease is not significant, however, the relative error associated with this method
in the estimates of rainfall rate becomes large since the fluctuation of surface cross section,
which remains finite even when there is no rain, translates to the absolute error of the rain
estimates. In order to avoid inaccuracies in the attenuation correction when rain is weak, a
hybrid of the surface reference method and the Hitschfeld-Bordan method is used [Iguchi and
Meneghini, 1994]. In versions 5 and 6 of 2A25, the errors in these methods are treated in a
probabilistic manner. Because the relationship between the error in α and that in the
Hitschfeld-Bordan method changes substantially with the attenuation, the relative weight on
the surface reference method to the Hitschfeld-Bordan method varies with the attenuation.
When rain is very weak and the attenuation estimate is small, the PIA estimate from the
surface reference is effectively neglected. With the introduction of the hybrid method, the
divergence associated with the Hitschfeld-Bordan method is also prevented.
When the PIA estimate from the surface reference ( .σ 0) is unavailable, it is replaced by an
equivalent .σ 0c that would make the attenuation-corrected Z-profile near the surface nearly
constant vertically if the correction by the surface reference method is applied with this
equivalent .σ 0c. This does not imply that the final vertical profile near the surface after the
attenuation correction becomes constant because of the use of the hybrid method. Note also
31
________________________________________________________________________
that negative values of .σ 0c are reset to zero. The use of .σ 0c instead of .σ 0 from the surface
reference seldom occurs in V6. The attenuation correction procedure requires two processing
cycles. In the first cycle, the correction is made without taking the attenuation by cloud liquid
water (CLW), water vapor (WV) and molecular oxygen (O2) into account. From the
attenuation-corrected profile, the rainfall rate at the surface is estimated. Based on this
rainfall rate and the statistical relationship between the surface rain rate and the vertical
profile of cloud liquid water, the attenuation of radar rain echo caused by CLW is estimated
at each range bin. Similarly, the attenuation due to WV is estimated from the estimated
surface temperature and by assuming the 90% relative humidity within the raining footprint
and 70% outside the raining area. The attenuation due to O2 is a simple function of the
altitude. Then in the second cycle, the vertical profile of Zm is corrected for the attenuation by
CLW, WV and O2, and this attenuation-corrected Zm is corrected for the attenuation by rain.
The corrections for the non-uniform beam filling effect in the attenuation correction and the
conversion from Ze to rainfall rate are not made in V6, although the non-uniformity of rain
distribution, i.e., the low resolution variability of the PIA for a given angle bin is calculated
from the PIAs at the angle bin in question and the eight surrounding angle bins in V6 as well
as in V5. The rainfall estimates are calculated from the attenuation-corrected Ze -profiles by
using a power law: R=aZeb in which the parameters a and b are both functions of the rain
type, existence of bright-band, freezing height, storm height and absolute height. Effects of
the difference in the raindrop size distribution by rain type, the phase state, the temperature,
and the difference in terminal velocity due to changes in the air density with height are taken
into account. The parameters a and b are expressed as a function of the adjustment parameter
( ε) of α in the k-Ze relation and adjusted in accordance with the α-adjustment in the
attenuation correction. The final estimate of R is obtained as the expectation of all-possible
values of R with the probability p( ε).
- Input data
Input files:
1C-21 HDF data file
2A-21 HDF data file
2A-23 HDF data file
Input swath data from 1C-21 which are used in 2A-25:
binClutterFreeBottom[][]
binEllipsoid[]
binStormHeight[][]
binSurfPeak[]
geolocation[][]
minEchoFlag[]
normalSample[][]
scanStatus.dataQuality
scanStatus.missing
scanStatus.prStatus2
scanTime
scLocalZenith[]
32
________________________________________________________________________
scRange[]
Input swath data from 2A-21 which are used in 2A-25:

pathAtten[]
reliabFlag[]
reliabFactor[]
sigmaZero[]
Input swath data from 2A-23 which are used in 2A-25:

rainType[]
warmRain[]
status[]
freezH[]
HBB[]
Input header data from 1C21 used in 2A-25:

rayHdr[].rayStart
rayHdr[].mainlobeEdge
rayHdr[].sidelobeRange[]
Output data
Output files:
2A-25 HDF data file
VI file
Output data (in alphabetical order):
attenParmAlpha[][] k-Z parameter alpha at 5 nodes

attenParmBeta[] k-Z parameter beta
correctZFactor[][] attenuation-corrected Z factor in dBZ
epsilon[] correction factor with the hybrid method
epsilon_0[] correction factor with the SRT
errorRain[] error estimate of rain rate near surface in dB
errorZ[] error estimate of Z near surface in dB
e_SurfRain[] estimated rain rate at the actual surface
freezH[] freezing height from 2A23
geolocation[][] geolocation
method[] method used
navigate navigation data
nearSurfRain[] estimated rain rate near surface
nearSurfZ[] estimated Z near surface
nubfCorrectFactor[][] non-uniform beam filling correction factors
ParmNode[][] bin numbers of 5 nodes for alpha, a and b
pia[][] path-integrated attenuations from final Z,
in surface clutter, and from 2A21
33
________________________________________________________________________
precipWaterParmA[][] PWC-Z parameter a in PWC=a*Z^b at 5 nodes

precipWaterParmB[][] PWC-Z parameter b in PWC=a*Z^b at 5 nodes
precipWaterSum[] sum of PWC from rain top to surface
qualityFlag[] quality flag

rain[][] rainfall rate in mm/h.
rainAve[][] average rainfall rate between 2 and 4 km
rainFlag[] status flag for rainfall estimate
rainType[] rain type from 2A23
rangeBinNum[][] bin numbers of BB, storm top, etc.
reliab[][] reliability of the output
scanStatus scanStatus
scanTime scanTime
scLocalZenith[] spacecraft local zenith angle
sigmaZero[] surface scattering cross section sigmaZero from 2A21
spare[][] spare
thickThPIZ[] range bin number where PIZ > threshPIZ
weightW[] weight for the calculation of epsilonf
xi[][] normalized standard deviation of PIA
zeta[][] integral of alpha*Zm^beta
zeta_mn[][] mean of zeta over 3x3 IFOVs
zeta_sd[][] standard deviation of zeta
zmmax[] maximum of Zm
ZRParmA[][] Z-R parameter a in R=a*Z^b at 5 nodes
ZRParmB[][] Z-R parameter b in R=a*Z^b at 5 nodes
Interfaces with other algorithms

As described in "Input data" section, 2A25 reads data from 1C21, 2A21 and 2A23.
The output data of 2A25 is used in 3A25 and 3A26.
3 .5 TRMM Data format
3 .5 .1 Introduction to HDF
The Hierarchical Data Format, or HDF, is a multi object file format for sharing scientific
data in a distributed environment. HDF was created at the National Center for
Supercomputing Applications to serve the needs of diverse groups of scientists working on
projects in various fields. HDF was designed to address many requirements for storing
scientific data, including:
• Support for the types of data and metadata commonly used by scientists.
• Efficient storage of and access to large data sets.
• Platform independence.
• Extensibility for future enhancements and compatibility with other standard formats.
34
________________________________________________________________________
In this document, the term HDF data structures will be used to describe the primary
constructs HDF provides to store data. These constructs include raster image, palette,
scientific data set, annotation, vdata, and vgroup. Note that the construct vgroup is designed
for the purpose of grouping HDF data structures. HDF files are self-describing. The term
“self-description” means that, for each HDF data structure in a file, there is comprehensive
information about the data and its location in the file. This information is often referred to as
metadata. Also, many types of data can be included within an HDF file. For example, it is
possible to store symbolic, numerical and graphical data within an HDF file by using
appropriate HDF data structures.
Figure 3.6 HDF Data Structure
HDF can be viewed as several interactive levels. At its lowest level, HDF is a physical file
format for storing scientific data. At its highest level, HDF is a collection of utilities and
applications for manipulating, viewing, and analyzing data stored in HDF files. Between
these levels, HDF is a software library that provides high-level and low-level programming
interfaces. It also includes supporting software that makes it easy to store, retrieve, visualize,
analyze, and manage data in HDF files. The basic interface layer, or the low-level API, is
reserved for software developers. It was designed for direct file I/O of data streams, error
handling, memory management, and physical storage. It is a software toolkit for experienced
HDF programmers who wish to make HDF do something more than what is currently
available through the higher-level interfaces. Low-level routines are available only in C.
35
________________________________________________________________________
The HDF application programming interfaces, or APIs, include several independent sets of
routines, with each set specifically designed to simplify the process of storing and accessing
one type of data. Although each interface requires programming, all the low-level details can
be ignored. In most cases, all one must do is make the correct function call at the correct
time, and the interface will take care of the rest. Most HDF interface routines are available in
both FORTRAN-77 and C.
3 .5 .2 Primary HDF Platforms
The HDF library and utilities are maintained on a number of different machines and
operating systems.
A lists the primary machines and operating systems HDF is ported to.
Figure 3.7 Primary HDF Platform
For a complete list of the machines, operating systems (with versions), C and FORTRAN-
77 compilers (also with versions), refer to the file named “INSTALL” in the root directory
of the HDF distribution.
3 .6 Methodology
The whole methodology can be divided into different stages:
1. Input data collection:
For the Input data collection, TRMM mission index file downloaded for the respective
year. This TRMM mission index file is mainly used to find out unusually intense storms
in regions with frequent rainfall where it is difficult to decide which storm is stronger.
Detailed discussion about TRMM mission index is given in paragraph (3.6.1).
2. Data processing:
In this stage downloaded data processed using C and FORTRAN programming. Detailed
description about the using software & the programming is discussed in paragraph
(3.6.3.1.1,3.6.3.1.2).
3. Output
In this stage, final output is generated in ASCII format shown in Table4.1.
36
________________________________________________________________________
The complete process is shown in a flow process diagram (please see figure 3.8)
Figure 3.8 Flow Process Diagram
37
________________________________________________________________________
3 .6 .1 Software
For the present study, various software like TSDIS Orbit viewer GrADS1.9 and HDF
Library are used. Briefly discussed about the software are presented below.
3.6.1.1 TSDIS Orbit viewer
The Orbit Viewer for TRMM data was developed by NASA. The Orbit Viewer is written in
the IDL language, which is manufactured by RSI [http://www.rsinc.com].
3.6.1.1.1 Freeware Installation under Microsoft Windows
Currently, the only version of Microsoft Windows that the Orbit Viewer is tested on is
Windows XP. An easy way to obtain the Orbit Viewer is by downloading the Zip file from
the TSDIS web site [ http://tsdis.gsfc.nasa.gov]. Place the Zip file in the top folder of a hard
disk of Microsoft Windows system. If the Zip file ends in *.exe, it is a self-extracting Zip
file. To unzip it, double-click it. If the Zip file ends in *.zip, use a compression program such
as WinZip to unzip it. Either way, when the Zip file is uncompressed, it created the orbit
folder. Go into the orbit folder and double-click on the setupWIN.exe program to set up the
Orbit Viewer. Virus protection software may warn you about the two Visual Basic Scripts
that the setup program runs. These scripts create a shortcut for the Orbit Viewer on desktop,
but even if that fails you can still use the Orbit Viewer.
3.6.1.1.2 Data Display
The Orbit Viewer provides many ways to display TRMM data, Arrays are emphasized
because most TRMM data is stored in arrays rather than tables or text fields. Basic Display
3.6.1.1.2.1 Array
The most frequently made images in the Orbit Viewer are high-resolution horizontal images
of data arrays. The following steps lead through the process of opening a file, choosing a data
array, and choosing a zoom region. After installation of software, double-click the Orbit
Viewer icon or the TSDISorbitViewer\orbitWIN.exe program and then used the File Open
menu item to open a data file. The main window of the Orbit Viewer will exist throughout
Orbit Viewer session (figure 3.9). For this example, opened a 2A25 file, which contains the
three-dimensional rainfall rate estimated from observations of the TRMM Precipitation
Radar (PR). Now that, opened a file in the Orbit Viewer, selected an array from the list on
the left side of the main window. In figure 3.12, chosen the “rain” array. From here on, this
array referred to as “2A25.rain” because the 3D rainfall rate estimate is named “rain” in
2A25 files. If the array has more than two dimensions, use the slider on the left side of the
window to choose one element of the additional dimension. In this example, the rain array
has 80 vertical levels going from the highest sampling level (height 1) down to the earth’s
ellipsoid (height 80). Here selected height 70, which corresponds to a height of 2.5 km. Now
that chosen a variable and a vertical height, selected a zoom region by clicking somewhere
on the global image at the bottom of the Orbit Viewer’s main window. Using the Option
Zoom Resolution menu item, increase the magnification by switching from the default zoom
38
________________________________________________________________________
image size of 20 degrees to 5 or 10 degrees. To improve the image quality, selected Option
Interpolate_ Data. If clicked on a data point in the zoom image, detailed information about
that observation will be printed in the lower-right corner of the window. Once created a zoom
image in the main window of the Orbit Viewer, it becomes possible to open several
additional windows to examine the array in different ways. Selected the View Array menu
item to see what windows are available for the currently displayed array. For example,
chosen View Array Histogram to open a histogram window. A few of the View Array menu
items become active even before selected a zoom region. For example, created a statistical
summary of all the arrays in the file by selecting View Array Quick_File_Summary.
Figure 3.9 The Main Window of TSDIS orbit viewer
39
________________________________________________________________________
3.6.1.1.2.2 Table
A table is a data structure with rows and columns, and each column can have a different data
type, such as integer, float, or string. In TRMM HDF files, tables are stored as Vertex Data
(Vdata). Most TRMM files contain two to five tables. Most tables in swath files (level 1 or 2
files) have 10 to 20 columns and one row for each scan in the swath. To get a list of the
tables in a particular file, selected the View Table menu item. Selected a table, the Table
Viewer window appeared (figure 3.10). Selected the name of a column from the list on the
left side of the Table Viewer window. A plot of that whole column appeared, along with a
detail plot of a portion of it. To change the resolution of the detail plot used the pull-down
menu at the bottom of the window. To change which portions of the column appears in the
detail plot; used the slider at the bottom of the window. To save the plot to an image file,
clicked the Save button.
Figure 3.10 The Table Viewer
3.6.1.1.2.3 Text Field
Metadata and occasionally others sorts of information are stored as text fields in TRMM
HDF files. To view the metadata of a file, selected the View Metadata menu item, and the
Metadata window appeared (figure 3.14). Due to a peculiarity in the way TRMM HDF files
40
________________________________________________________________________
are structured, other text fields, such as the Grid Structure field, are stored as Vdata Tables
and can therefore be accessed by selecting the View Table menu item.
Figure 3.11 Meta data
3.6.1.1.2.4 Vertical View
Swath arrays have two horizontal dimensions: scan number and field-of-view (fov). Some
swath arrays have a third dimension that represent height above the earth’s surface, method
of calculation, or some other quantity. One of the strengths of the TRMM dataset is the
ability to resolve the three-dimensional structure of rainfall rate, so the Orbit Viewer provides
two ways to make vertical cross sections of swath arrays: a fixed cross section and a variable
cross section. The first method holds one dimension of the swath fixed and makes an image
along the other two dimensions. In contrast, the second method cuts the swath at any azimuth
angle.
3.6.1.1.2.5 Fixed Cross Section
Selected a three-dimensional array, such as the 2A25.rain array, in the main window of the
Orbit Viewer and chosen a zoom region by clicking in the global image at the bottom of the
main window. Selected the View Array Fixed_Cross_Section menu item, and a new window
appeared (figure 3.12). In the Fixed Cross-Section window, click on the left-hand column to
chosen which type of cross section to make: horizontal, along track, or across track.
41
________________________________________________________________________
Whichever cross section chosen, one dimension of the array fixed. To change the displayed
element of the fixed dimension, move the slider at the bottom of the Fixed Cross Section
window. Click on the image to open a detail window.
Figure 3.12 Fixed cross section
3.6.1.1.2.6 Variable Cross Section
Selected the three-dimensional array in the main window of the Orbit Viewer and chosen a
zoom region by clicking in the global image at the bottom of the Orbit Viewer’s main
window. Selected the View Array Variable_Cross_Section menu item in the main window,
and a new window appeared (figure 3.13). Chosen the geographic endpoints of the cross
section either type them into the text fields or click the button labeled “select endpoints with
the cursor.” If chose to select the endpoints with the cursor, click on the zoom image where
the cross section should begin, hold down the mouse button, and drag the mouse to where
would like the cross section to end. Then, let go of the mouse button. The Orbit Viewer
draws a white line where the cross section would lie. Click the button labeled “display the
cross section.”
42
________________________________________________________________________
Figure 3.13 Variable Cross Section
3.6.1.1.3 3D View
Visualized the full 3D structure of an array, it has to be helpful to examine a 3D surface. The
Orbit Viewer constructed a 3D surface that contains all data above the user-supplied
threshold. The following steps described how to generate a 3D surface from a swath array
using the Orbit Viewer. Opened a 2A25 file and selected the Precipitation Radar’s “rain”
array (in mm/h) from the left-hand column of the Orbit Viewer’s main window.
Alternatively, opened a 2A12 file and select “precipitable water,” (in g/m 3) which is
estimated from TMI observations. Whatever array chosen, click on the global image at the
bottom of the Orbit Viewer’s main window to select a zoom region. Selected the View Array
3D_Viewer menu item from the main window. A small window titled “Select Volume”
appeared to allow to select the volume to be displayed three dimensionally (figure 3.14).
Also, a white box appeared on the zoom image. Moved the box to so that it covers the data
that visualize. To do this, clicked on the “+” symbol in the middle of the box, hold down the
43
________________________________________________________________________
mouse button, dragged the mouse to the center of the region of interest, and then let go of the
mouse button. Alternatively, the center of the 3D region fixed by typing its latitude and
longitude values into the Select Volume window. Selected the size of the region in degrees of
latitude, use the pull-down menu.
Figure 3.14 Select 3D Locations
Once chosen the region of interest, click the Select Data button. After a brief delay, a 3D
Control window and a 3D Image window appeared (figure 3.18). The sliders in the 3D
Control window allowed changing the observation angle (1) and the angle of the light source
(2). Figure 3.8 shows the definition of these angles. Another slider allowed the threshold to
be changed that defines what data are included inside the 3D surface (3). All data values
greater than that threshold are inside the 3D surface. This slider is label in the units of the
data being displayed. In figure 3.18, the threshold is 1.5 mm/h (15 * 0.1 mm/h). Zoomed in,
use the “Zoom” slider (4). Once zoomed in, that the portion of the data that interests, fallen
outside the image. Solved this problem, move the image using the horizontal or vertical shift
sliders (5).
44
________________________________________________________________________
Figure 3.15 3D control window (top), and 3D Image window (bottom)
45
________________________________________________________________________
Figure 3.16 Definition of the sliders in the 3D Control window
3.6.1.1.4 Generic View
By selected the View Array Generic View menu, three windows appeared: a Control window
that determines what portion of the array is displayed and the Image and Text windows
where the array is displayed.
46
________________________________________________________________________
Figure 3.17 Generic View Control Window
In this example, we examined the 2A25.rain array. The HDF file contains names for the
dimensions of this array, namely height, ray, and scan. The height is the vertical dimension,
the ray is the field-of-view, and the scan is the number of the scan along the satellite track.
The check boxes next to each dimension determine which dimensions are currently being
displayed. In this case, field-of-view and scan number are being displayed, i.e. this is a
horizontal slice at height bin 70. This corresponds to a height of 2.5 km above the earth’s
ellipsoid (80 - 70 = 10 cells from the ellipsoid; 1 cell is 0.25 km of altitude). By default, the
text values are displayed with the full precision of a four-byte floating-point number, i.e. an
IDL character format of “g13.6”. In order to see a larger portion of the array as text, It has
been reduced the precision of the displayed value. To do this, select the File Change Format
menu item from the Generic View Text window. If you choose the three-digit integer format
(“i3”) and expand the window by stretching its lower-right corner, you will get the text
shown below in figure (3.18).
47
________________________________________________________________________
Figure 3.18 The Generic View Text window
3.6.1.1.4.1 Array Summaries
There are two ways to summarized an array in the Orbit Viewer. It created a histogram of an
individual array or a summary of all the arrays in a file. Selected the View Array Histogram
menu item to created a histogram. (figure 3.19).
Figure 3.19 Histogram window
48
________________________________________________________________________
Selected the View Array File_Summary menu item to created a file summary. Wait while the
Orbit Viewer reads all the arrays in the file and calculates statistics then appropriate result
has come,a new window that contains a list of all the arrays and their statistics. Below is
shown the statistics for the Rain variable in a 2A25 subset file (figure 3.20).
Figure 3.20 A portion of a file summary

3.6.1.1.5 Data Search
3.6.1.1.5.1 TRMM Mission Index
The TRMM Mission Index helped, to find unusually intense storms in regions with frequent
rainfall where it is difficult to decide which storm is stronger. The Mission Index can also
helped to find storms in semi-arid regions even though rainfall is rare there. Once determined
a TRMM orbit of interest using the Mission Index, It have enough information to order that
orbit’s full-resolution file from the TRMM archive. Full resolution files can be obtained from
the Goddard DAAC at [http://daac.gsfc.nasa.gov] .
The TRMM Mission Index summarized the more than 10 terabytes of TRMM Microwave
Imager (TMI) and Precipitation Radar (PR) single-orbit rainfall products produced since
January 1998. Based on the 3G68 algorithm, the Mission Index consists of daily average
rainfall rate on a 1-degree grid with an overlay of the satellite track and orbit numbers. Each
Mission Index file summarized the single-orbit files for one year for one of the following two
surface rainfall algorithms: the TMI 2A12 algorithm and the PR 2A25 and the PR 2A25
algorithm. Each file is small, only 11 megabytes, for easy downloading. The first step is to
download a Mission Index file from the TSDIS web site [http://tsdis.gsfc.nasa.gov]. Opened
the file in the Orbit Viewer and selected the avg_rain_rate array. Next, clicked on the global
image to select a zoom region. At this point, the Orbit Viewer calculate a time series of
49
________________________________________________________________________
rainfall rate for that year in that zoom region. By clicking on a peak in the time series, the
scientist created a low-resolution image of heavy rainfall on a particular day. For refine the
region in the time series, used the File Select_Region menu item in the Time Series window.
In the Select Region window that appeared, type in the latitude and longitude boundaries of
the region of interest and click the Apply button. Alternatively, click on the box to the left of
the words “Select region with cursor.” Then, rubber bands a region in the zoomed image of
the Orbit Viewer’s main window. Once chosen a region, the time series is recalculated
automatically.
Figure 3.21TRMM Mission index window
50
________________________________________________________________________
Figure 3.22 Daily Time Series Window
Figure 3.23 Using the TRMM Mission Index to Find Data of Interest
3 .6 .2 GrADS1.9 b4 Linux software
GrADS is an abbreviation for "Grid Analysis and Display System". GrADS is a data analysis
and visualization language created by Brian Doty of the Center for Ocean-Land-Atmosphere
51
________________________________________________________________________
Studies (COLA). GrADS was originally designed to work with gridded four-dimensional
arrays with dimensions of latitude, longitude, height, and time. Additional types of data can
be read by recent versions of GrADS. The GrADS software and documentation can be
obtained from the COLA web site: [http://grads.iges.org]. [http://grads.iges.org]. For UNIX
platform, download a compressed tar file that contains executables for your type of UNIX.
Uncompress and Un tar the distribution file. Then set the GADDIR environment variable to
point to the grads/dat directory of the distribution. Started a GrADS command line on a
UNIX computer, go into the grads/bin directory and type "grads" at the UNIX command
prompt. The up arrow key allowed to edit previous commands that issued and then reissue
them.. If used the Orbit Viewer to create a GrAD compatible data file that contains the rain
field from a TRMM 2A25 file obtained a *.dat GrADS binary data file and a *.ctl GrADS
control file. When created the file in the Orbit Viewer, keep the filename to 19 characters or
less, because there appears to be a limit to how long filenames can be in GrADS. Depending
on what directory working in, ". Than used the following commands on the GrADS
command line to display the contents of this pair of *.dat and *.ctl files.
GrADS> open 2A25.floyd.ctl
GrADS> level 70 70
GrADS> display data
GrADS> wi image.gif
The file line opened the file. In this example, the 2A25.rain field has 80 levels going from the
oceans surface around level 80 to 20 km above the ocean at level 0. Level 70 frequently has
rain, so the "level 70 70" command selects level 70. The Orbit Viewer stores data in an array
called "data" when the Orbit Viewer creates a GrADS data file. The "display data" command
creates an image of level 70 of the data array. To write the image on the screen to a GIF
image on hard disk, use the "wi" command. Instead of typing commands on the GrADS
command line, it is sometimes helpful to type then into a text file and then run that text file
in the GrADS command line.
GrADS Commands
-ga> open pr-050115.40853.ctl
Scanning description file: pr-050115.40853.ctl
Data file pr-050115.40853.grd is open as file 1
LON set to -164.95 164.95
LAT set to -34.95 34.95
LEV set to 1 1
Time values set: 1999:7:1:0 1999:7:1:0
ga> set display color white
ga> c
ga> set gxout grfill
ga> set lat 0 40
LAT set to 0 40
ga> set lon 60 100
LON set to 60 100
ga> d r1
Contouring: 0 to 8 interval 1
ga> cbarn
ga> quit
52
________________________________________________________________________
Figure 3.24 Rainfall shown in GrADS
3 .6 .3 HDF Library
The HDF library is a collection of software routines that provides two type of interfaces
which allow the user to work with HDF file . A brief capsule describing these interfaces is
provided below:
1.Low –Level Interface
2.Application Programming Interfaces (APIs)
Component of the HDF library include the base library, the multi – file library, the jpeg
library, and the gzip library. The HDF library also provides a set of command – line utilities
that allow the user to work with HDF files outside of the interfaces.
3.6.3.1 Programming languages supporting the HDF library
53
________________________________________________________________________
As of the current release of HDF ,the only programming languages which are supported by
the HDF library are C and FORTRAN. Although the HDF library code is only written in C,
the library provides both a FORTRAN and java Interface which converts the code to C and
allow the user to call the HDF routines.
3.6.3.1.1 GGG.C PROGRAMMING
/*************************************************************************
*
* Last modified February 10, 1999
*
* Program to read in a TRMM HDF-EOS data file, and write out
* user-selected SDS arrays and Vdata tables as separate flat binary
* files. Metadata (global attributes) are written out in a separate
* ascii file with extension .meta
*
* Data in output files will be the same number type (float32, int32,
* int16, int8, uint8) as the corresponding array in the input (HDF)
* file.
*
*
* USAGE :
*
* read_trmm <trmm_hdf_file>
*
*
* NOTES :
*
* 1) Please make sure that the buffer arrays are dimensioned properly
* via the PARAMETER statements
*
* 2) Output filenames automatically created by replacing 'HDF' suffix
* of input file with selected parameter (SDS) names
*
**************************************************************************/
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <hdf.h>
#define FIELD_SIZE 1024 /* maximum length of all the field names */

#define MAX_DIMS 20 /* maximum number of dimensions in sds */
#define MAX_SDS 200 /* maximum number of sds */
#define MAX_VDATA 200 /* maximum number of vdata */
#define IN_TAIL0 "hdf" /* possible input filename extension */
#define IN_TAIL1 "HDF" /* possible input filename extension */
54
________________________________________________________________________
main(int argv, char *argc[])

{
FILE *meta_fp, *sds_fp, *vd_fp;
int32 sd_id, sds_id, dim_id, file_id;

int32 index, attr_index, sds_idx, dim_index;
int32 nindex, sds_index[MAX_SDS];
int32 rank, attributes, n_datasets, n_file_attrs, num_type, count;

int32 n_records, nparm, temparm, cal_data_type, num_element ;
int32 status, i,j, ifield;
int32 dim_sizes[MAX_VAR_DIMS];
int32 start[MAX_DIMS], stride[MAX_DIMS], edges[MAX_DIMS];
int32 interlace, vdata_size, field_index, field_size, field_order;

int32 field_type, vdata_idx, vdata_ref, vdata_id;
int32 vdata_index[MAX_VDATA], total_vref[MAX_VDATA], total_vidx=0;
char fields[FIELD_SIZE], attr_name[64], sds_name[64];
uint8 *vdatabuf; /* buffer to retrieve the vdata data */

char vdata_name[VSNAMELENMAX]; /* buffer to retrieve the vdata name */
char *str, *attr_str, *fieldname;

char *infile, *tempfile, *nstr;
char outfile[120];
float64 gain, gain_err, offset, offset_err;
char8 *c8;
int8 *i8;
uint8 *ui8;
int16 *i16;
uint16 *ui16;
int32 *i32;
uint32 *ui32;
float *f32;
float64 *f64;
/*--------------------------------------------------------------------*/
55
________________________________________________________________________
if (argv != 2) {
printf("Usage: read_trmm <trmm_hdf_file> \n\n");
exit(-1);
}
if (Hishdf(argc[1]) == 0) {
printf("Error: %s is not a valid HDF file, or file not found\n\n", argc[1]);
exit(-1);
}
/*
* Open connections to original (input) TRMM HDF file
*/
file_id = Hopen (argc[1], DFACC_READ, 0);
if ( file_id == -1 ) {
printf("Error opening input file : %s\n", argc[1]);
exit(-1);
}
/*
* Open HDF SDS and Vdata Interfaces
*/
sd_id = SDstart(argc[1], DFACC_READ);
if (sd_id == -1) {
printf ("SDstart failed.\n");
HEprint (stdout, 0);
exit (-1);
}
status = Vstart (file_id);

if (status == -1 ) {
printf ("Vstart failed.\n");
exit (-1);
}
/*
Generate the prefix for output file
*/
infile = strdup(argc[1]);
nstr = strstr(infile, IN_TAIL0);
if (nstr == NULL)
nstr = strstr(infile, IN_TAIL1);
if(nstr != NULL)
strcpy(nstr, "");
else
strcat(infile, ".");
56
________________________________________________________________________
/*
* Obtain number of SDSs and Global Attributes in input file
*/
status = SDfileinfo(sd_id, &n_datasets, &n_file_attrs);

if (status != 0 ) {
printf("SDfileinfo failed. \n");
exit (-1);
}
/*
* List out SDSs contained in input HDF file
*/
printf("\n Number of SDS arrays in file: %d\n\n",n_datasets);
for (index = 0; index < n_datasets; index++) {

sds_id = SDselect(sd_id, index);
status = SDgetinfo(sds_id, sds_name, &rank, dim_sizes, &num_type, &attributes);

if (status != 0) {
printf ("SDgetingo failed.\n");
exit (-1);
}
printf(" %2d) %-25s dimensions = ", index+1, sds_name);

for(dim_index = 0; dim_index < rank; dim_index++)
printf("%7d", dim_sizes[dim_index]);
printf(" \n");
status = SDendaccess(sds_id);
}
/*
* Prompt user for desired parameters to be written as binary file
*/
printf("\n Enter total number of parameters to write out or 0 to exit program\n");
scanf("%d", &nparm);
if(nparm == 0) {
nindex = 0;
sds_index[nindex] = 0;
}
else if (nparm == n_datasets){
nindex = 0;
for (index = 0; index < n_datasets; index++)
sds_index[nindex++] = index+1;
}
57
________________________________________________________________________
else {
printf("\n Enter parameter numbers from above list separated by white space\n");
nindex = 0;
for (index = 0; index < nparm; index++) {
scanf("%d", &temparm);
if(temparm <= n_datasets)
sds_index[nindex++] = temparm;
}
}
if(nparm > 0) {
printf(" The valid parameter numbers you Enter are: ");
for (index = 0; index < nindex; index++)
printf(" %d ", sds_index[index]);
}
else
printf(" No sds selected -- # of parameters = 0 \n");
/*
* Now loop through SDSs and write selected array(s) to binary files
*/
if(nindex > 0) {
for (index = 0; index < nindex; index++) {
num_element = 1;
sds_idx = sds_index[index]-1;
sds_id = SDselect (sd_id, sds_idx);
status = SDgetinfo(sds_id, sds_name, &rank, dim_sizes, &num_type, &attributes);
status = SDgetcal(sds_id, &gain, &gain_err, &offset, &offset_err, &cal_data_type );
printf("\n ****************************************\n");
printf(" SDS name = %s\n", sds_name);
printf(" SDS type = %4d\n", num_type);
printf(" SDS rank = %4d\n", rank);
printf(" SDS scale = %f\n", gain);
printf(" SDS dims = ");
for (j = 0; j < rank; j++) {
printf("%6d", dim_sizes[j]);
num_element *= dim_sizes[j];
}
printf ("\n");
for (j = 0; j < rank; j++) {

edges[j] = dim_sizes[j];
stride[j] = 1;
start[j] = 0;
}
/*
* Read TRMM sds arrays
58
________________________________________________________________________
*/
switch (num_type)
{
case DFNT_FLOAT32: f32 = (float *) malloc( num_element * sizeof(float)); break;
case DFNT_FLOAT64: f64 = (float64 *) malloc( num_element * sizeof(float64));
break;
case DFNT_INT8: i8 = (int8 *) malloc( num_element * sizeof(int8)); break;
case DFNT_UINT8: ui8 = (uint8 *) malloc( num_element * sizeof(uint8)); break;
case DFNT_CHAR8: c8 = (char8 *) malloc( num_element * sizeof(char8)); break;
default: printf("not valid data type\n"); break;
}
switch (num_type)
{
case DFNT_FLOAT32: status = SDreaddata (sds_id, start, NULL, edges, f32); break;
case DFNT_FLOAT64: status = SDreaddata (sds_id, start, NULL, edges, f64); break;
case DFNT_INT8: status = SDreaddata (sds_id, start, NULL, edges, i8); break;
case DFNT_UINT8: status = SDreaddata (sds_id, start, NULL, edges, ui8); break;
case DFNT_CHAR8: status = SDreaddata (sds_id, start, NULL, edges, c8); break;
}
if ( status != 0 )
printf("\n SDreaddata failed on data set %s. \n", sds_name);
/*
* Out put sds to binary data file
*/
tempfile = strdup(infile);
strcpy(outfile, tempfile);
strcat(outfile, sds_name);
sds_fp = fopen(outfile,"wb");
if (sds_fp == NULL)
printf("Open output sds file failed. \n");
printf("sds file %s. \n", outfile);
switch (num_type)
{
case DFNT_FLOAT32: status = fwrite(f32, num_element * sizeof(float),1,sds_fp);
break;
59
________________________________________________________________________
case DFNT_FLOAT64: status = fwrite(f64, num_element * sizeof(float64),1,sds_fp);

break;
case DFNT_INT8: status = fwrite(i8, num_element * sizeof(int8),1,sds_fp); break;
case DFNT_UINT8: status = fwrite(ui8, num_element * sizeof(uint8),1,sds_fp);
break;
case DFNT_INT16: status = fwrite(i16, num_element * sizeof(int16),1,sds_fp);
break;
break;
case DFNT_INT32: status = fwrite(i32, num_element * sizeof(int32),1,sds_fp);
break;
break;
case DFNT_CHAR8: status = fwrite(c8, num_element * sizeof(c8),1,sds_fp); break;
}
if ( status == 0 )
printf("\n fwrite failed on data set %s. \n", sds_name);
else
printf("\n Binary file successfully created : %s\n", outfile);
fclose(sds_fp);
strcpy(outfile, "");
free(tempfile);
switch (num_type)
{
case DFNT_FLOAT32: free(f32); break;
case DFNT_FLOAT64: free(f64); break;
case DFNT_INT8: free(i8); break;
case DFNT_UINT8: free(ui8); break;
case DFNT_CHAR8: free(c8); break;
}
status = SDendaccess(sds_id);
} /* end of for loop for sds */
} /* end of if statement for nindex */
status = SDend(sd_id);
printf(" \n\n");
/*
* Now read in the Vdata tables and write non-dimension Vdatas to
* separate output files
*/
vdata_ref = - 1;
60
________________________________________________________________________
vdata_ref = VSgetid(file_id, vdata_ref);

if ( vdata_ref == -1 ) {
printf("\n No Vdatas found in data set\n");
exit(-1);
}
printf("\n\n List of Vdata name: (rec_num rec_size)\n\n");

while(vdata_ref != -1) {
vdata_id = VSattach(file_id, vdata_ref, "r");
status = VSinquire(vdata_id, &n_records, &interlace,

fields, &vdata_size, vdata_name);
if ( status != 0 ) {
printf("\n VSinquire failed on vdata %s\n", vdata_name);
exit(-1);
}
if ( strstr(vdata_name, "fakeDim") == NULL &&

strstr(vdata_name, "offset") == NULL &&
strstr(vdata_name, "factor") == NULL ) {
total_vref[total_vidx++] = vdata_ref;
printf(" %2d) %-30s %5d %5d \n", total_vidx, vdata_name,
n_records, vdata_size);
}
status = VSdetach( vdata_id );
vdata_ref = VSgetid ( file_id, vdata_ref );
}
/*
* Prompt user for desired vdata to be written as binary file
*/
printf("\n Enter total number of vdata to write out or 0 to exit program\n");
scanf("%d", &nparm);
if(nparm == 0) {
nindex = 0;
vdata_index[nindex] = 0;
}
else if (nparm == total_vidx){
nindex = 0;
for (index = 0; index < total_vidx; index++)
vdata_index[nindex++] = index+1;
}
else {
printf("\n Enter parameter numbers from above list separated by white space\n");
nindex = 0;
for (index = 0; index < nparm; index++) {
61
________________________________________________________________________
scanf("%d", &temparm);
if(temparm <= total_vidx)
vdata_index[nindex++] = temparm;
}
}
if(nparm > 0) {
printf(" The valid vdata numbers you Enter are: ");
for (index = 0; index < nindex; index++)
printf(" %d ", vdata_index[index]);
}
else
printf(" No vdata selected -- # of vdata = 0 \n");
}
/*
*/
3.6.3.1.2 FORTRAN PROGRAMMING
CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC
C
C PROGRAMME TO FIND THE AREAL RAINRATE
C FROM RADAR DATA
C
C type5 is real*4, type6 is real*8
C type20 is integer*1, type22 is integer*2
C rank implies number of dimensions and scale
C implies the factor by which the data has to
C be multiplied for extraction.
implicit none
real*4 surf_rain(49),geo(2,49)
real*8 scanTime
real*4 lat,lon,xmax,xmin,ymax,ymin
real*4 g,missing
integer*4 hour,minut,sec,i,j,nx,ny
real,allocatable:: sumf(:,:),avalf(:,:)
integer,allocatable:: icountf(:,:)
integer ixcol,iycol,izcol,ncol
character orbit*12,infile*19
print*, 'ENTER ORBIT NUMBER:'

read(*,1)orbit
1 format(12a)
c ---- Opening geolocation file --------------------
62
________________________________________________________________________
open(30,file='/home/cold/sunita
12A25.'//orbit//'.6.02B.02B.geolocation',status='old',access=
1'direct',form='unformatted',recl=49*2*4)
c ---- Opening surfacerain file -------------------
open(31,file='/home/COLTD/SUNITA/
12A25.'//orbit//'.6.02B.02B.nearSurfRain',status='old',access=
1'direct',form='unformatted',recl=49*4)
c ---- opening scan time file ----------------------
12A25.'//orbit//'.6.02B.02B.scan_time',status='old',access=
1'direct',form='unformatted',recl=8)
c ---- opening the output ascii file for pr data ---
open(15,file='/home/COLTD/SUNITA/pr-'//orbit//'.dat')
c------------------------------------------------------
c ---- Reading each file record by record, here each---
c ---- record corresponds to a scan line of PR --------
c ---- there are 9247 scan lines in one pass of PR ----
c------------------------------------------------------
xmax=95.0
xmin=65.0
ymax=40.0
ymin=0.0
c xmax=90.3
c xmin=79.7
c ymax=24.2
c ymin=12.8
i=1
100 read(30,rec=i,err=99)geo
read(31,rec=i)surf_rain
read(32,rec=i)scanTime
i=i+1
c ---------------------------------------------------
c ---- Total time is given in seconds, converting it
c ---- into hour, minut and seconds ------------------
63
________________________________________________________________________
c ---------------------------------------------------
hour=scanTime/60/60
minut=(scanTime-hour*3600)/60
sec=(scanTime-hour*3600-minut*60)
c -----------------------------------------------------
c ----- Each scan line has 49 observations i.e. pixels-
c -----------------------------------------------------
do j=1,49
lat=geo(1,j)
lon=geo(2,j)
c ----------------------------------------------------
c ---- Extracting the data from region of intrest ----
c ----------------------------------------------------
if(lat.ge.ymin.and.lon.ge.xmin
1 .and.lat.le.ymax.and.lon.le.
1 xmax)write(15,9)hour,minut,sec,
1 lat,lon,surf_rain(j)
enddo
9 format(3(i2.2,1x),3(f9.3,1x))
goto 100
99 continue
close(15)
c ----------------------------------------------------
c ---- Defining variables for the average subroutine--
c ----------------------------------------------------
g=0.1
missing=-999.99
ncol=6 !Total No. Of Columns in File.
ixcol=5 !Longitude column No.
iycol=4 !Latitude column No.
izcol=6 !Z Column No.
c ----------------------------------------------------
c ----- calculating total no. of grids along lon and -
c ----- lat directions ------------------------------
c ----------------------------------------------------
nx=int((xmax-xmin)/g)+1
ny=int((ymax-ymin)/g)+1
64
________________________________________________________________________
infile='pr-'//orbit//'.dat'
c ----------------------------------------------------
c - Allocating the memory spaces for arrayed variables
c ----------------------------------------------------
allocate(sumf(nx,ny))
allocate(icountf(nx,ny))
allocate(avalf(nx,ny))
c ----------------------------------------------------
c ---- Calling the subroutine to grid the pr data ----
c ----------------------------------------------------
call average(xmin,xmax,ymin,ymax,ncol,ixcol,nx,ny
1 ,iycol,izcol,infile,g,missing,sumf,icountf,avalf)
c ----------------------------------------------------
c Writing the output of subroutine to grads grd file
c ----------------------------------------------------
1pr-'//orbit//'.grd',status='unknown', form='unformatted'
1,access='direct',recl=nx*ny*4)
write(16,rec=1)((avalf(i,j),i=1,nx),j=1,ny)
deallocate (sumf,icountf,avalf)
close(16)
c ----------------------------------------------------
c ---- writing the corresponding grads control file --
c ----------------------------------------------------
1pr-'//orbit//'.ctl',status='unknown')
write(17,12)"DSET pr-",orbit,".grd"
write(17,15)'undef ',missing
write(17,1)'TITLE TRMM RADAR RAIN RATE'
write(17,10)'XDEF',nx,'LINEAR',(xmin+g/2.),g
write(17,10)'YDEF',ny,'LINEAR',(ymin+g/2.),g
write(17,1)'ZDEF 1 LINEAR 1 1'
write(17,1)'TDEF 1 LINEAR 00:00Z01JUL1999 1dy'
write(17,1)'VARS 1'
65
________________________________________________________________________
write(17,1)'r1 0 99 PR Rain'
write(17,1)'ENDVARS'
c ----------------------------------------------------
c ---- writing the corresponding grads script file --
c ----------------------------------------------------
1pr-'//orbit//'.gs',status='unknown')
write(18,1)"'reinit'"
write(18,16)"'open pr-",orbit,".ctl","'"
write(18,1)"'set gxout grfill'"
write(18,1)"'set display color white'"
write(18,1)"'c'"
write(18,2)"'set lat",ymin,ymax,"'"
write(18,2)"'set lon",xmin,xmax,"'"
write(18,1)"'set grads off'"
write(18,1)"'set mpdset hires'"
write(18,1)"'d r1'"
write(18,1)"'run cbarn'"
write(18,13)"'draw title PR rainrate(mm/h)",orbit,"'"
2 format(a,1x,f7.2,1x,f7.2,a)
12 format(8a,12a,4a)
16 format(8a,12a,4a,a)
14 format(a,2x,a)
15 format(a,2x,f9.2)
10 format(a,1x,i4,1x,a,1x,f10.3,1x,f5.3)
13 format(a,1x,12a,a)
end
c ------------------------------------------------------------
c ------------------------------------------------------------
c ----------- FORTRAN CODE OF SUBROUTINE AVERAGE -------------
c ------------------------------------------------------------
c ------------------------------------------------------------
SUBROUTINE AVERAGE(lonmin,lonmax,latmin,latmax,ncol,
+ ixcol,nx,ny,iycol,izcol,infile,g,missing,sum,icount,
+ aval)
character infile*19
dimension sum(nx,ny),icount(nx,ny),aval(nx,ny),a(ncol)
real latmax,latmin,lonmin,lonmax,missing
66
________________________________________________________________________
real alon,alat,val,x,y
sum=0.0
icount=0
aval=0.0
open(unit=1,status='old',file='/home/COLTD/ SUNITA/'//infile)
c ------------------------------------------------------------
c -- reading from file and assigning lat, lon and parameter --
c ------------------------------------------------------------
5 read(1,*,err=51,end=99)(a(i),i=1,ncol)
alon=a(ixcol)
alat=a(iycol)
val=a(izcol)
if(alon.lt.lonmin.or.alon.gt.lonmax.or.alat.lt.latmin
1 .or.alat.gt.latmax) goto 5
if(val.lt.0.0) goto 5
c ------------------------------------------------------------
c ---- main gridding starts ----------------------------------
c ------------------------------------------------------------
x=lonmin
do j1=1,nx
x=x+g
if(alon.le.x) goto 21
enddo
21 y=latmin
do j2=1,ny
y=y+g
if(alat.le.y) goto 22
enddo
22 sum(j1,j2)=sum(j1,j2)+val
icount(j1,j2)=icount(j1,j2)+1
51 goto 5
67
________________________________________________________________________
99 do j2=1,ny
do j1=1,nx
if (icount(j1,j2).eq.0) then
aval(j1,j2)=missing
else
aval(j1,j2)=sum(j1,j2)/icount(j1,j2)
endif
enddo
enddo
return
end
68
________________________________________________________________________
CHAPTER 4
4 RESULT AND VALIDATION
4 .1 General Discussion
In the present study one task taken up to highlight the use of remote sensing as tool along
with appropriate software to study Rainfall Estimation. Present study is based on the TRMM
Satellite Mission.
Here with the help of some free downloaded software and programming, extract the rainfall
data from TRMM-PR sensor and convert it into ASCII format (table 4.1). This rainfall data is
more accurate than other data such as rain gauge, GOES data etc. TRMM Satellite cover all
over the world, To better understand global changes and their mechanisms, scientists require
long-term observations of the Earth System and the interactions of the atmosphere, oceans,
and lands. In particular, the hydrological cycle and energy budget are among the most
important issues to be studied to understand the global climate system.
More than two-thirds of global precipitation falls in tropics and subtropics. The latent heat
released by this rainfall plays a vital role in the global atmospheric circulation and the energy
budget. The major objective of TRMM is to increase our understanding of global climate
change through space-based measurements of tropical rainfall and Earth radiation. TRMM is
a key element of "Mission to Planet Earth," an international science program to provide the
measurements necessary to understand our Planet as a system.
The rain radar, with its first space-based flight on TRMM, will be a prototype for future
space borne microwave remote sensors. The non-sun synchronous TRMM orbit will provide
observations of the diurnal cycle of rainfall, which are not possible from existing sun-
synchronous satellites. Results from TRMM will build the foundation for continuing
programs for long-term observations of global rainfall and climate change.
Using this TRMM data we can able to get the rainfall data for any specified area of interest.
This rainfall data is mainly useful for regional as well as micro level planning with out
accessing those areas, which are difficult to traverse to collect the rain gauge measurements.
The data would be analyzed with the available software instantaneously to predict natural
hazards like instant floods, how to mitigate the damages. The resultant output data from
TRMM satellite, of any area would be in the format of ASCII as shown in table 4.1.
69
________________________________________________________________________
Hour Min Sec Latitude Longitude Surface rain

mm/hr
23 48 14 22.459 81.322 0.00
23 48 14 22.496 81.354 0.00
23 48 14 22.533 81.376 2.00
23 48 14 22.570 81.398 3.00
23 48 14 22.606 81.419 0.00
23 48 14 22.642 81.441 4.00
23 48 14 22.678 81.462 0.00
23 48 14 22.714 81.488 0.00
23 48 14 22.749 81.504 0.00
Table 4.1 Surface Rainfall
4 .2 Validation
Validation is the process of checking if something satisfies a certain criterion. In present

study TRMM normalized rainfall data was validated with respect to IMD (Indian
Meteorological Department) data. IMD rainfall data was taken from RMSI in-house available
data backup.
Normalized TRMM Rainfall data is extracted from TSDIS orbit viewer software using
TRMM mission index file 2004. for this validation purpose another TRMM data of Solapur
district, Maharastra, India, was used. The extent of area is given below:
Study Area- Solapur

Latitude – 17. 40
Longitude - 75. 54
70
________________________________________________________________________
There are following steps which are used in data validation study.
1. Extraction of Normalized daily rainfall data of respective year (2004) from TRMM
mission index.
2. Calculation of mean monthly rainfall data from TRMM.
3. Plotting of best fit curves between TRMM and IMD rainfall data.
The mean monthly derived rainfall data from TRMM and IMD are given below in table 5.1.
Month TRMM IMD

(Mean Monthly) (Mean Monthly)
Jan 16.1 0
Feb 1.1 0
Mar 3.5 0
Apr 12.5 0
May 28.6 0
Jun 177.9 204.8
Jul 447.4 413.2
Aug 91.9 66.8
Sep 392.9 462.4
Oct 41.1 27.2
Nov 16.5 0
Dec 8.8 0
Table 4.2 shows mean monthly rainfall data of TRMM and IMD
71
________________________________________________________________________
Figure 4.1 show the relation between TRMM and IMD mean monthly rainfall data
4 .2 .1 Conclusion
The above table 5.1 shows that TRMM rainfall indicates some rain in Pre monsoon period
(Jan to April). But IMD data shows no rain at that time. That is because PR sensor observes
some precipitation in clouds at pre monsoon time but that is not perceptible. That why in
IMD data shows ZERO rainfall at Pre monsoon period. But in the case of monsoon period
both TRMM and IMD data shows good correlation.
4 .2 .2 Future scope
Since the TRMM produces rainfall data pixel wise which is already in interpolated form
(spatial form like theisson polygon). Future scope of work include, extraction of rainfall data
equivalent to an area covered by one IMD rain gauge station. Using the total number of pixel
present in an area equivalent to IMD area, the mean rainfall data would be collected and will
be validated against available IMD data.
72
________________________________________________________________________
CHAPTER 5
5 CONCLUSION
In present study TRMM data was used to compute rainfall for an area covering Orissa state,
India. TRMM provides rain fall data, over any specified area of interest. In this study TRMM
data was validated with respect to IMD data .The result was 46% of confidence. The best fit
was obtained with higher order polynomial order (3rd). This indicates that there are some
data anomalies within both the datasets. Gap in IMD data and measurement of very fine
atmospheric wet condition by TRMM produces data anomaly. Precipitation Radar emits
radar pulses toward Earth, which are then reflected by precipitation particles back to the
radar. By measuring the strength of the returned pulses, the radar estimates rainfall rates.
According to PR rainfall capturing technology the rainfall is estimated from all clouds type
and water vapor content in the atmospheric zone, but in actual scenario all these clouds are
not perceptible. At this time IMD Produce Rainfall as a zero. This case creates less
confidence level while performing regression.
The major conclusions of the study:

 Estimates rainfall from all clouds type and water vapor content in the atmospheric
zone
 Rainfall is estimated along the path of satellite pass.
 It estimated small atmospheric moisture variation which is non-perceptible producing

data mismatch with IMD.
 In this study the maximum confidence level gained was 46% with respect to IMD
data.
73
________________________________________________________________________
LIST OF WEBSITES VISITED

1 http://trmm.gsfc.nasa.gov /
2.http://daac.gsfc.nasa.gov/precipitation/TRMM_REA
3.http://daac.gsfc.nasa.gov/precipitation/TRMM_README/TRMM_3A46_readme.shtm
4.http://grads.iges.org/grads/
5.http://grads.iges.org/grads/downloads.html
6. http://tsdis02.nascom.nasa.gov/tsdis/Documents/ICSVol4.pdf
7.ftp://ftp-tsdis.gsfc.nasa.gov/pub/TSDISorbitViewer/download
8.http://tsdis.gsfc.nasa.gov/tsdis/TSDISorbitViewer/mission.htm
9.http://www.eoc.jaxa.jp/satellite/satdata/trmm_e.html
10.http://www.eorc.nasda.go.jp/TRMM/document/orbitviewer/index.htm
74
________________________________________________________________________
BIBLIOGRAPHY:
 Wilheit, T. T., A. T. C. Chang, and L. S. Chiu, 1991: “Retrieval of Monthly Rainfall
Indices from microwave Radiometric Measurements Using Probability Distribution
Functions,” J. Atoms. Oceanic. Techno., Vol. 8, pp. 118-136.
 Chen, W. J. and C.-C. Li, 2000: “Oceanic Rain Rate Retrievals Using TRMM
Microwave Imager Multi-Channel Brightness Temperatures During the 1998
SCSMEX,” Terrestrial, Atmospheric and Oceanic Sciences, Vol. 11, pp. 765-788.
 Lovejoy, S., and G. L. Austin, 1979: The delineation of rain areas from visible and IR
satellite data from GATE and mid-latitudes. Atoms.-Ocean, 17, 77-92.
 Manobianco, J., S. Koch, V. M. Karyampudi, and A. J. Negri, 1994: The impact of
assimilating satellite-derived precipitation rates on numerical simulations of the
ERICA IOP 4 cyclone Mon. Wea. Rev., 122, 341-365.
 Marécal, V., and J.-F. Mahfouf, 2000: Variational retrieval of temperature and
humidity profiles from TRMM precipitation data. Mon. Wea. Rev., 128, 3853–
3866.
 Marécal, V., É. Gérard, J.-F. Mahfouf, and P. Bauer, 2000: The comparative impact
of the assimilation of SSM/I and TMI brightness temperatures in the ECMWF 4D-
Var system. Q J. R. Meteorol. Soc., 127, 1123-1142
 Meneghini, R., 1998, J. Appl. Meteor., 37, 924-938.
 Meneghini, R., and J. Jones, 1993, J. Appl. Meteor., 32, 386-398.
 Short, D.A., K. Shimizu, B. Kedem, 1993, J. Appl. Meteor., 32, 182-192.
 Kedem, B., L.S. Chiu, G.R. North, 1990, J. Geophys. Res., 96, 1965-1972.
 Atlas, D., D. Rosenfeld, D.A. Short, 1990, J. Geophys. Res., 95, 2153-2160.
 Meneghini, R., J.A. Jones, T. Iguchi, K. Okamoto, and J. Kwiatkowski, 2001, J.
Appl. Meteor., 40, 568-585.
 For background information about TRMM, see the December 2000 special issues of
the Journal of Climate and the Journal of Applied Meteorology. See also the TRMM
Home Page: [http://trmm.gsfc.nasa.gov]. TRMM data provided by NASA / NASDA /
CRL.
 Owen Kelley, John Stout, and Menas Kafatos. "Contentbased Browsing of Data from
the Tropical Rainfall Measuring Mission (TRMM)." Proceedings of the IEEE
Scientific and Statistical Database Conference (SSDBM). George Mason University.
Fairfax, Va. July 18-20, 2001. [3] http://tsdis.gsfc.nasa.gov/trmmopen
 Keiffer, H. H., and R. L. Wildey, 1996: Establishing the moon as a spectral radiance
standard. J. Atmos. Oceanic Tech., 13, 360-375. Kumagai, H., T. Kozu, M. Satake, H
. Hanado and K. Okamoto, 1995: Development of an active radar calibrator for the
TRMM precipitation radar. IEEE Trans. Geosci. Remote Sens., GE-33, 1316-1318.
 Meneghini, R. and T. Kozu, 1990. Spaceborne Weather Radar. Artech House, 199 pp.
Okamoto, K. and T. Kozu, 1993. TRMM precipitation radar algorithms. Proc
. IGARSS'93, 426-428. Satake, M., K. Oshimura, Y. Ishido, S. Kawase and T. Kozu,
1995. TRMM PR data processing and calibration to be performed by NASDA.
Proc. IGARSS'95. 57-59
75

M.SC Project Report

Uploaded by

Copyright:

Available Formats

You might also like

M.SC Project Report

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

M.SC Project Report

Uploaded by

Copyright:

Available Formats

Rainfall Estimation by Remote Sensing

Flood and drought monitoring, management of agricultural crops, forecasting hydroelectric

I feel it as a special responsibility to convey my special thanks to my thesis supervisor

I wish to express my heartfelt thanks to Mr. Rajesh Kumar, Research Assistant,

I wish to express my deep sense of gratitude to my parents and special thanks to my

Figure 1.1 Various types of platforms in Remote Sensing .................................................. 9

Table 1.1 subdivision of optical wave length region ................................................................ 9

1 .1 Introduction of Remote sensing

1 .2 Types of Remote Sensing

In respect to Wavelength Regions

 Visible and Reflective Infrared Remote Sensing

1 .3 Bands Used in Remote Sensing

1 .4 Spectral Characteristics vis-à-vis Different Systems

Figure 1.1 Various types of platforms in Remote Sensing

Name Wavelength (mm)

Table 1.1 subdivision of optical wave length region

Figure 1.2 The electromagnetic spectrum

1 .5 Energy Interactions, Spectral Reflectance and Colour Readability in Satellite Imagery

1 .5 .2 Spectral reflectance and color readability

1 .6 Planck radiation law

h = 6.625X 10-27 erg-sec (Planck Constant)

1. Rainfall estimation by remote sensing technique using TRMM satellite data.

2. Validation of TRMM data with respect to IMD

2 .2 Rain Gauge- Measurement

2 .3 WSR-88D Precipitation Products

2 .5 The Operational GOES Infrared Rainfall Estimation

2 .6 Tropical Rainfall Measuring Mission (TRMM)

3 .1 General Introduction About TRMM

Decision-makers and researchers require accurate spatial and temporal precipitation

Launch weight Approx. 3.62 ton

Launcher H-II Rocket

Shape At lift-off: 5.1 m (length), 3.7 m

Weight Fuel: 890 kg

Power Approx. 1100 W [Ave. 850 W]

Attitude control Zero momentum three-axis

Design life 3 years and 2 months

Table 3.1 Main Characteristics of the TRMM Satellite

Figure 3.1 TRMM orbit path

Figure 3.2 TRMM Orbit paths

3 .3 .1 TRMM Microwave Imager (TMI)

3 .3 .2 Precipitation Radar (PR)

3 .3 .3 Visible Infrared Scanner (VIRS)

3 .3 .4 Cloud and Earth’s Radiant Energy System (CERES)

3 .3 .5 Lightning Imaging Sensor (LIS)

Figure 3.3 Shown TRMM sensor scanning system

3 .4 TRMM precipitation radar system description

Figure 3.4 Measurement Concept Of Precipitation Radar

The main performances of the precipitation radar are as follows:

1) Frequency : 13.796 GHz (f1) and 13.802 GHz (f2) (2 frequencies,

9) Scan width : During observation: 215 km or more (Between centers of

includes logarithm detector and A/D conversion ) to the

33) Scan angle bin

a) Number of scan angle : 103

Table 3.2 Major Parameters of TRMM PR

PRODUCT NO NAME CONTACT PRODUCTS ALGORITHM

1B21 PR calibration JAXA/EOC Total received Conversion of the

2A21 Surface R. Meneghini Path integrated Estimation of path

FILE meta_fp, sds_fp, *vd_fp;

uint8 vdatabuf; / buffer to retrieve the vdata data */

char str, attr_str, *fieldname;