TRMM LBA Rain Gauge/Disdrometer Draft Operations Plan
Please direct any questions, comments, or suggestions to Larry Carey at
carey@radarmet.atmos.colostate.edu
I. Rain Gauges/ Disdrometers
1. Introduction
The purpose of the rain gauge and disdrometer deployment during
TRMM-LBA is 1) to obtain independent measurements of surface rain rate
for calibration and comparison with radar derived estimates in a
variety of rainfall regimes and at various ranges from the two radars,
2) to investigate the temporal and spatial variability of rainfall in
Rondonia, and 3) to determine how this variability impacts the accuracy
of the ground validation (GV) radar products.
In addition to scientific goals, there are also important
logistical issues. The chosen site locations must be 1) available
private land, 2) easily accessible during the wet season (e.g.,
preferably on paved vs. dirt roads if possible), and 3) reasonably
close to one of the two radars for purposes of data collection and
instrument maintenance. In addition, each of the 2-D video
disdrometers (2DVD) requires shelter and air conditioning for their
accompanying PCs. The rest of the instruments utilize data loggers
which are designed to withstand the elements. To accomplish these
scientific and logistical goals, the rain gauges and disdrometers
listed in Sec. I1 above will be deployed at five separate locations in
the vicinity of Ji Parana.
2. Shipping and Deployment Plan
Thirty Qualimetric tipping bucket rain gauges, and 30 UNIDATA data
loggers will be air shipped from NASA Goddard Space Flight Center to
Ji Parana for installation in November 1998. One 2DVD, one Joss impact
disdrometer, and three APL impact disdrometers with supporting
instrumentation have been included for shipment in the NOAA / profiler
seatainer which will be shipped to Ji Parana in October 1998. A
second 2DVD will be air shipped to Ji Parana from the University of
Iowa in late November or early December. Six additional APL impact
disdrometers will be shipped to Ji Parana at approximately the same time.
Routing of the air shipments will depend on time and cost requirements,
unless otherwise dictated because of import licensing requirements.
Final instrument locations and exact network configurations will
depend on the results of a final site survey to be conducted prior to
instrument set-up. The final site survey will be conducted during
October 1998 and instrument set-up is expected to occur primarily
during November 1998. During the final site survey, necessary site
preparation will occur. For example, wooden platforms will be
constructed on which the rain gauges and disdrometers will be placed.
If necessary, fencing will be constructed to keep cattle away from the
instruments. The final set-up will involve the physical placement,
programming, and calibration of the gauges and disdrometers according
to TRMM procedures. It is anticipated that the rain gauges and lower
maintenance disdrometers (APL and Joss) will begin data collection by
1 Dec 98. The 2-D video disdrometer will require more elaborate
set-up, calibration, and maintenance and therefore will likely be
deployed at the beginning of the IOP (e.g., first week of January 1999).
One 2-D video and an APL disdrometer will be operated at the TOGA radar site (Abracos Hill). A
location at intermediate distances (30 - 40 km) from both of the radars
would have been ideal from a scientific and data analysis perspective.
However, two logistical issues took precedence in this siting
location: 1) The accompanying data processing and display PC will be
housed in one of the TOGA seatainers. If the instrument is not sited
at the TOGA radar, then a shelter with air conditioning will have to be
identified. 2) Collocation with the TOGA radar will allow radar
scientists to maintain and monitor the 2DVD carefully and frequently.
Otherwise, a remote site causes logistical and manpower issues in
arranging frequent site visits. The drop size distribution (DSD)
data from this disdrometer will be compared to the S-pol radar
observations located approximately 62 km to the southeast. Given the
choice between the two radars, the TOGA seemed to be the obvious site
choice. The S-pol is a 10-cm (minimal attenuation problems),
polarimetric radar while the TOGA radar is at an attenuating wavelength
(5 cm) and is non-polarimetric. The rest of the instrumentation will
be deployed at 4 separate networks (Fig. 1).
Fig. 1 Updated road map containing the approximate locations
of the four rain gauge networks relative to the S-pol (S) and TOGA
(T) radars, major roads, and cities and towns. This color map will
be updated for the final operations plan.
Although rain gauge data will be used as a reference for the radar
estimation of rainfall, it is critical to establish first the sampling
error of the gauges. The design and proposed locations of the first
two gauge/disdrometer networks has this goal in mind. The
configuration of these networks is optimized to study the spatial
variability of tropical rainfall at a variety of distance classes from
10 meters to 4 km, permitting the estimation of the shape of the
spatial correlation function of rainfall. This information can then be
utilized to estimate the gauge sampling error throughout the entire
rain gauge deployment (given the local gauge configuration, number of
gauges, and the individual rain gauge measurement error).
The triangle design of the rain gauge cluster utilized in networks
#1 and #2 was developed to capture the small scale variability of
rainfall with the minimal number of gauges. There is a need to
understand the small scale variability of rainfall to reduce the errors
in the radar-rainfall estimation. Radar-rainfall errors are a
combination of radar errors due to sampling, noise of the system,
incorrect calibration, and natural variability of rainfall. If the
spatial correlation of rainfall can be determined for distances
corresponding to the size of a radar-rainfall grid (ie. 2 km), the
natural variability of rainfall can be removed from the radar-rainfall
error. To determine the spatial correlation function, gauges must be
placed over a range of intergauge distances, but there must be a number
gauges at the same distance to improve the sample statistics. A
triangle cluster allows for both. For networks # 1 and 2, there will be three samples at each
vertices of the triangle. Then there will be at least 3 more samples
at each vertex of a sub triangle of gauges placed in the large
network (See Figs. 2 and 3). Ideally, the gauges should deployed at distances over powers
of 10 so scaling of rainfall can also be examined. For instance,
gauges could be located at ranges of 1m, 10 m, 100 m, 1 km, out to 2
km. Finally, the remaining gauges should be dispersed outside the
rain gauge cluster to increase the sample area.
Fig. 2 Idealization of gauge/disdrometer network # 1 arranged in two
nested triangles. There are a total of 12 Qualimetric rain gauges and 6
disdrometers. See the text for more details.
Network #1 (Fig. 2) will consist of twelve rain gauges and four
disdrometers (one 2DVD, one Joss, and two APL) arranged in two nested,
equilateral triangles with sides of 2 and 4 km. The inner (2km)
equilateral triangle will contain nine rain gauges and all four of the
disdrometers. Two vertices of the inner triangle will each contain a
tight cluster of 3 rain gauges which will be separated by approximately
10 to 100 meters. As shown in Fig. 1, one of these two vertices will
be collocated with the profiler at the Ji Parana airport which is
situated approximately 30 km (50 km) to the north-northeast
(east-southeast) of the S-pol (TOGA) radar. The Joss, one 2-D video,
and one APL disdrometer will be located at the profiler site. The
purpose of collocating the three disdrometers is to provide
cross-validation between the three instruments. The PC used to process
and display the 2-D video disdrometer data will be housed in one of the
profiler seatainers. The second vertex possessing 3 closely clustered
gauges will be accompanied by an APL disdrometer. The third vertex of
the inner triangle will contain a single rain gauge and an APL disdrometer. Along two sides
of the inner triangle, a rain gauge will placed approximately 0.5 km
from a vertex. Finally, the outer (4 km) equilateral triangle will
consist of one rain gauge at each vertex with an APL collocated at one of the vertices. This multi-instrumented
(profilers, disdrometers, gauges) ground validation network should
provide unprecedented observations of DSD, rainfall rate, and
precipitation vertical structure in tropical continental rainfall and
will be crucial to achieve improved understanding of the ground radar (both
polarimetric and Doppler) measurements and the TRMM satellite measurements.
Fig. 3 Idealization of gauge/disdrometer network # 2 arranged in two
nested triangles similar to network #1. There are a total of 12 Qualimetric
rain gauges and 4 disdrometers. The primary difference between the layout
of the two networks is the number and location of disdrometers. The gauge
layout is the same. See the text for more details.
Network # 2 (Fig. 3) will be centered approximately 55 km to the
north-northwest of the S-pol radar along BR 364 (Fig. 1) and will
consist of twelve rain gauges and three APL disdrometers arranged in a
similar nested triangle configuration (e.g., equilateral triangles
measuring 2 and 4 km on a side). One APL disdrometer will be placed
with a rain gauge at each of the three vertices of the inner triangle.
There will also be an APL disdrometer at one of the vertices of the outer triangle.
The purpose of separating 24 gauges and 10 disdrometers into two
separate networks is to maximize the probability of capturing a
significant rainfall event on any given day while still allowing the
investigation of rainfall spatial variability on small scales.
Fig. 4 Idealization of gauge network # 3 and 4 arranged in a triangle
configuration. See the text for more details.
Rain gauge networks # 3 and #4 (Fig. 4) will be located 80 and 120
km to the northwest of the S-pol radar along BR 364 as shown in Fig. 1.
Each of these networks will contain three gauges arranged in a
triangle measuring approximately 1 to 1.5 km on a side. The purpose of
these two networks is to provide information regarding the range
behavior of the radar estimation of rainfall at distances far from
S-pol, which is critical for the development of TRMM GV rainmap products.
The combination of the four gauge networks provide reference data for the S-pol
radar at ranges from 30 to 120 km . In addition, reference data will
be available at 20 to 50 km in range to both the southeast and
northwest of the TOGA radar. The gauges will be clustered at scales
less than a typical radar rainfall grid (say 2 km) so that a
representative rain gauge sample average over the corresponding grid
point can be compared to the radar rainfall estimate with confidence.
Of equal importance, each of the 4 networks will be located along a
well travelled, fast moving, and paved road (BR 364). This should
facilitate data collection and instrument maintenance.
3. Staffing plan
The tipping bucket rain gauges and the APL and Joss disdrometers are
stand alone, low maintenance equipment which require only periodic data
collection and maintenance visits. As a result, site visits to each
gauge or impact disdrometer will typically occur once a week to download
data and check on instrument status. These visits will be the responsibility
of the "rain gauge-disdrometer scientist (RGDS), who will be designated by
the operations director. The RGDS will be responsible for site visits to the
gauge/disdrometer networks, necessary maintenance, data collection, and
quality control in the field. Based on available manpower, the RGDS may
utilize personnel from other platforms in order to complete the site
visits.
Based on TRMM's experience during TEFLUN-B, the 2-D video
disdrometer is a more delicate and higher maintenance instrument when
operating 24 hours a day in a high humidity environment. The RGDS will
also be responsible for monitoring and servicing these instruments as needed.
Similar to the rain gauge and impact disdrometer tasks, the RGDS is
expected to utilize assistance from scientists dedicated to other platforms
(e.g., TOGA or S-POL radar crew members) to assist in the maintenance and
monitoring of the 2DVDS. Preferably, this scientist should have prior field
experience with the 2DVD. This individual can coordinate data monitoring,
data collection, QC efforts, and any necessary maintenance.
4. Operations scenario
The data loggers for the Qualimetrics rain gauges have enough
memory to record approximately 1500 mm of rainfall. The expected total
IOP (January and February) rain accumulation at gauges in Rondonia is
from 200 to 1300 mm. However, it is advisable to visit the sites
as frequently as possible to check on instrument status.
Therefore, all rain gauges and their collocated APL disdrometers should
be checked and their data down loaded and saved with a laptop PC
approximately once a week. Because the gauges and disdrometers are
distributed over a relatively large area and because accessibility
may be difficult in the wet season, it is expected that the RGDS will
spend each day collecting data from a small subset of the gauges and
APL disdrometers. Over the course of a week, it is anticipated that all
of the gauge/disdrometer sites can be visited. As stated above, the RGDS
may utilize additional personnel to perform the gauge-disdrometer site
visits. Thus, the composition of the gauge/disdrometer site visit teams
will be determined in cooperation between the RGDS, the operations
director, and the chief radar scientists at the S-pol and TOGA radars.
The 2-D video disdrometers will require more frequent attention.
Based on recent experience in a hostile tropical environment during TEFLUN-B,
site visits should occur on a daily basis at both 2DVD deployment locations.
The frequency of these site visits during the experiment will depend on the
performance of the instruments in the field and the availability of personnel
for the task.
5. Data plan
a. Collection method
Rain gauge and APL disdrometer data will be collected from
the UNIDATA data loggers using a portable lap top
computer and customized software provided by the TRMM Project Office.
This procedure will be accomplished once a week. Three (3)
back-up copies of the gauge and disdrometer data should be made
immediately on a PC floppy disk or tape drive. In addition, three (3)
back up disks or tapes should be made of the 2-D video disdrometer data
approximately once every week. The Joss disdrometer data will be automatically
archived on a computer housed within the profiler seatainer. Back-up copies
of this data will be made on Jaz drive media by personnel responsible for
maintaining the profiler equipment.
b. QC plan
Quality control (QC) of the rain gauge and disdrometer data will
occur on a periodic basis during the IOP using data analysis software
on a laptop PC (or the 2DVD PC) provided by the TRMM Project Office,
manufacturers, and/or TRMM PIs. During the field campaign, data from
the gauges and APL and Joss disdrometers will be available for QC at
least once a week. Ideally, the in-field QC process would cover
all data from all instruments. However, it is recognized that the
volume of data combined with limitations in manpower will make complete
in-field QC a difficult goal to achieve. At a minimum, all data from
each gauge and disdrometer should be scanned and partially QC'ed in
order to identify obvious and serious instrument problems. Samples of
various gauge and disdrometer data should be QC'ed in complete detail.
The RGDS will be responsible for leading the in-field QC effort. Data
from the 2DVD should be viewed and analyzed on an "as needed" and
likely more frequent basis using software provided by the manufacturer
and/or TRMM PIs on the 2DVD PC.
Final, post-mission rain gauge and disdrometer data QC will be
accomplished by the TRMM office in conjunction with other TRMM LBA
investigators using the same procedures utilized for primary TRMM GV
site data.
c. Availability plan
The TRMM Project Office will make raw rain gauge and
disdrometer data gathered and periodically QC'ed in the field available
via FTP as soon as practical after conclusion of the IOP. Copies of all
data collected from the rain gauges will be provided to the assigned
Brazilian collaborating P.I. prior to departure of U.S. scientists.
d. Archive plan
The TRMM Project Office in collaboration with the TRMM TISDIS
should archive the raw and QC'ed rain gauge and disdrometer data on
mass store. In addition, back-up tape copies of raw and QC'ed data
should be maintained by the TRMM Project Office and TRMM PI's.