Level 4.2 (L4.2) processed upper-air sonde data

                Ambon, Biak, Jakartka-Cengkareng, Kupang, Makassar, Manado, Medan, Merauke, Padang, Palu, Pangkal Pinang, and Surabaya released: 27 May 2014

High vertical resolution (5-10m) sonde data that have been interpolated to uniform 5-hPa levels. Quality-control
flags are assigned objectively based on a number of QC checks and then adjusted by visual inspection.

upaqi_96035 - Medan corrected for RH/temperature errors
upaqi_96163 - Padang corrected for RH/temperature errors
upaqi_96237 - Pangkal Pinang corrected for RH/temperature errors
upaqi_96749 - Jakarta-Cengkareng corrected for RH/temperature errors
upaqi_96935 - Surabaya corrected for RH/temperature errors
upaqi_97014 - Menado corrected for RH/temperature errors
upaqi_97072 - Palu corrected for RH/temperature errors
upaqi_97180 - Makassar corrected for RH/temperature errors
upaqi_97372 - Kupang corrected for RH/temperature errors
upaqi_97560 - Biak corrected for RH/temperature errors
upaqi_97724 - Ambon corrected for RH/temperature errors
upaqi_97980 - Merauke corrected for RH/temperature errors


Indonesian Site Correction Description:

For Level 3.0 data, corrections of humidity have been applied with two
methods. One is a CDF matching based on the intercomparison of Vaisala RS92
and Meisei onboard the Mirai, and the second is a correction for the RH
discontinuity at the 0C level following a method provided by Sugidachi and
Fujiwara (2013).   Details can be found in Ciesielski et al. (2014). 

For the L3.1 data, in addition to the
corrections applied to produce Level 3.0 data, temperature discontinuity at
around 300hPa is corrected. Meisei radiosonde receiving system has a software
which reduces the influence of solar radiation reflected from the sensor
package, which is originally set to act from 300hPa level. Actually, it forces
the reduction of temperature at and above 300hPa. In particular, this feature
is found not only for daytime sounding but also for nighttime sounding,
indicating too much reduction. This is because the software reduces the
temperature regardless with actual condition and launch time, while the
soundings over the Indonesian Maritime Continent are typically conducted in
the local morning and evening time, when the errors caused by solar radiation
are expected to be small. Therefore, by assuming that temperature lapse rate
is close to the mean value between just below/above layers, the difference of
lapse rate from the mean is calculated, and the temperature difference deduced
from the lapse rate difference is added to the observed values (typical value
is about 0.2C). 

For the L3.2 data, in addition to the corrections applied to produce L3.0 and
L3.1 data, two additional corrections have been applied. At some Indonesian
sounding sites, the operators enter the wind data of 0m/s, 0deg as the initial
value regardless with actual wind condition. Thus, if the wind data of first
point, which is manually recorded by the operator, is 0m/s, 0deg, and any of
the first three data, which are the value measured by a Radiosonde, exceeds
0.5m/s, we judge the surface wind data is suspicious, and have replaced them
with 999m/s and 999deg, indicating missing values. Another correction scheme
is meant for reduction of erroneous heating/cooling data in daytime/nighttime
soundings. Since the observation was carried out on land, there are possible
that large heating/cooling may happen near the ground surface. However, we
found wrong cases. For example, while we confirmed a good agreement between
T/RH values measured using a calibrated handy sensor and that of radiosonde,
only surface meteorological data measured independently shows quite different
value. On the one hand, radiosonde sensor is exposed under the sun prior to
launch at some sites, which might cause sensor arm heating. This may also be
partly due to the launch site infrastructure condition, where the site is
surrounded by buildings. Therefore, near surface (surface ~ 40m) T/RH values
are replaced with the data obtained from linear extrapolation using above
layer data following Yoneyama et al. (2002).

Ciesielski, P. E., H. Yu, R. H. Johnson, K. Yoneyama, M. Katsumata, C. N.
Long, J. Wang, S. M. Loehrer, K. Young, S. F. Williams, W. Brown, J. Braun, T.
Van Hove, 2014: Quality-controlled upper-air sounding dataset for
DYNAMO/CINDY/AMIE: Development and corrections. J. Atmos. Oceanic Technol.,
31, 741-764.

Sugidachi, T., and M. Fujiwara, 2013: Corection of the stepwise change
observed at 0C in Meisei RS2-91, RS-01G, and RS-06G radiosonde relative
humidity profiles. J. Meteor. Soc. Japan, 91, 323-336.

Yoneyama, K., M. Hanyu, S. Sueyoshi, F. Yoshiura, and M. Katsumata 2002:
Radiosonde observation from the ship in the tropical region. JAMSTECR, 45,
31-39. Available from
"http://www.godac.jamstec.go.jp/catalog/data/doc_catalog/media/shiken45_04.pdf"

Questions or isssues with data should be referred to: Paul Ciesielski (paulc@atmos.colostate.edu)