OBSERVATIONS: Upper-Air Moisture Data Void Discovered . . . and Partially Rectified

Clough, C. T.; Padovani, Michael

doi:10.1175/2007bams2463.1

Cited by 1 publication

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Most early radiosonde hygrometers were considered unreliable under such cold conditions (Remsberg , and it was a standard practice until 1993 for U.S. stations to report humidity measurements as missing when T was below 2408C. This practice may have continued until at least 2006 at more than 200 radiosonde stations in South America, Australia, India, China, and other countries, even though some modern hygrometers are able to measure RH when T is as low as 21008C (Clough and Padovani 2008). To minimize the effect of this practice, McCarthy et al (2009) first rejected all the humidity reports when T , 2408C in .5% of a season's observations, and then they used the linear regression coefficients between monthly-mean T and RH, and between monthly-mean T and the natural logarithm of q in the 1995-2003 observations to estimate the mean RH and q of the missing (or rejected) observations from the mean of the associated T for all years.…”

Section: Known Discontinuities In Radiosonde Humidity Datamentioning

confidence: 99%

A New Approach to Homogenize Daily Radiosonde Humidity Data

et al. 2011

View full text Add to dashboard Cite

Radiosonde humidity records represent the only in situ observations of tropospheric water vapor content with multidecadal length and quasi-global coverage. However, their use has been hampered by ubiquitous and large discontinuities resulting from changes to instrumentation and observing practices. Here a new approach is developed to homogenize historical records of tropospheric (up to 100 hPa) dewpoint depression (DPD), the archived radiosonde humidity parameter. Two statistical tests are used to detect changepoints, which are most apparent in histograms and occurrence frequencies of the daily DPD: a variant of the Kolmogorov-Smirnov (K-S) test for changes in distributions and the penalized maximal F test (PMFred) for mean shifts in the occurrence frequency for different bins of DPD. These tests capture most of the apparent discontinuities in the daily DPD data, with an average of 8.6 changepoints (;1 changepoint per 5 yr) in each of the analyzed radiosonde records, which begin as early as the 1950s and ended in March 2009. Before applying breakpoint adjustments, artificial sampling effects are first adjusted by estimating missing DPD reports for cold (T , 2308C) and dry (DPD artificially set to 308C) conditions using empirical relationships at each station between the anomalies of air temperature and vapor pressure derived from recent observations when DPD reports are available under these conditions. Next, the samplingadjusted DPD is detrended separately for each of the 4-10 quantile categories and then adjusted using a quantile-matching algorithm so that the earlier segments have histograms comparable to that of the latest segment. Neither the changepoint detection nor the adjustment uses a reference series given the stability of the DPD series.Using this new approach, a homogenized global, twice-daily DPD dataset (available online at www.cgd.ucar.edu/cas/catalog/) is created for climate and other applications based on the Integrated Global Radiosonde Archive (IGRA) and two other data sources. The adjusted-daily DPD has much smaller and spatially more coherent trends during 1973-2008 than the raw data. It implies only small changes in relative humidity in the lower and middle troposphere. When combined with homogenized radiosonde temperature, other atmospheric humidity variables can be calculated, and these exhibit spatially more coherent trends than without the DPD homogenization. The DPD adjustment yields a different pattern of change in humidity parameters compared to the apparent trends from the raw data. The adjusted estimates show an increase in tropospheric water vapor globally.

show abstract

Section: Known Discontinuities In Radiosonde Humidity Datamentioning

confidence: 99%