The Challenges of Making Reference
Upper Air Measurements
Tom Gardiner
Emissions and Atmospheric Metrology Group,
National Physical Laboratory, UK
Ruud Dirksen
GRUAN Lead Centre, DWD Meteorological Observatory,
Lindenberg, Germany
RMetS meeting on Reference Observations and Calibration of
(Re)Analyses, ECMWF, 2nd July 2015
Welcome to the National Physical Laboratory
Contents
 Background to the National Physical Laboratory
 Motivation and goals of GRUAN
 Example GRUAN data product (Vaisala RS-92)
 Achievements and future developments
National Physical Laboratory





Founded in 1900
The UK’s National Measurement Institute
450+ specialists in Measurement Science
State-of-the-art laboratory facilities
The heart of the UK’s National Measurement System to
support business and society
 Maintains national measurement standards and
capabilities across a wide range of physical
measurements.
 Participates in international key comparisons and
supplementary comparisons
Emissions and Atmospheric
Metrology Group




30+ year history of environmental measurements
21 scientists, about half working in the field
Supported by consultants and students
Three main scientific areas :
 Emissions Monitoring
 Atmospheric Science and Climate Change
 Low Carbon Metrology
 EAMG is part of the Environmental Division at NPL
 The Division also covers :
 Gas and Particle Metrology
 Earth Observation and optical measurements
 The NPL Centre for Carbon Measurement
Motivation for GRUAN (GCOS
Reference Upper Air Network)
IPCC AR5 on long term water vapour trends :
Lower troposphere (PW):
“Radiosonde, GPS and satellite observations of tropospheric water vapor indicate
very likely increases at near global scales since the 1970s ….“
Upper troposphere:
“… the absence of a homogenized data set across multiple satellite platforms
presents some difficulty in documenting coherent trends from these records (of
upper tropospheric humidity).”
Stratosphere:
“Because of the large variability and relatively short time series, confidence in
long-term stratospheric H2O trends is low.”
Clear need for high quality reference measurements for climate
observations in the upper atmosphere.
Reference data needed to provide long term climate records and validation/QA
support for other data sources (satellites, operational networks).
Water vapor trends in the
troposphere?
e.g.: Lindenberg 8km (0:00 UT)
Water vapor trends in the
troposphere?
e.g.: Lindenberg 8km (0:00 UT)
Freiberg
RKS-2
RKS-5
MARZ
RS80
RS92
What is GRUAN?
The GCOS Reference Upper Air Network (GRUAN):
 Started in 2008.
 GRUAN is response to the need of WMO and the Global
Climate Observing System (GCOS) for the highest quality data
possible.
 Ground based network for reference upper air observations for
climate under GCOS and integrated into WIGOS.
 Currently 22 sites, with the aim to expand to 30 to 40 sites
worldwide.
 See www.gruan.org
GRUAN sites
GRUAN goals
 Maintain consistent observations
over decades
 Validation of satellite systems
 Numerical weather prediction
 Deliberate measurement
redundancy
 Standardization and traceability
 Quality management and
managed change
Priority 1:
•temperature
•water vapor
•pressure and wind
Priority 2: Ozone, …
Establishing reference quality
Uncertainty of
input data
Traceable
sensor
calibration
Transparent
processing
algorithm
Best estimate
+
Uncertainty
Black box
software
Proprietary
methods
Disregarded
systematic effects
What is needed to join GRUAN?
• A long-term continuous upper air measurement program with
proper change management.
• New system/software/procedure must be evaluated prior to
implementation.
• Quantification of systematic and random uncertainties.
• Verification by redundant observations (overlap).
• Collection of raw and meta data
Site certification
• Assessment of the site’s measurement program
• GRUAN-approved measurement quality
GRUAN data product for RS92
Good performance, 30% market share
Vaisala calibration of PTU sensors traceable to
SI standards
 Proprietary algorithms
GRUAN data product for RS92
Characterization & correction of instrument errors/biases
Well-documented
Vertically resolved error estimates
Error sources
• Temperature
• Radiation
• Humidity:
• Radiation
• Sensor time-lag
• Calibration
Radiation error: Laboratory
experiments
Shadow RS92 records background
temperature &Difference
ambient pressure
illuminated – background radiosonde
Simultaneous testing of 3 radiosondes
p=[3 hPa , ambient]
Radiation error correction model
Assumptions:
• On average 50% of
maximum insolation
• Radiative transfer
model for radiation
field (direct and
diffuse radiation)
• Climatological
clouds
Δ
T-correction profile
Sources of measurement uncertainty:
• Sensor orientation
• Ventilation
• Unknown radiation field (albedo)
• Lab measurements of the radiative
heating
• Ground check
• Calibration
RS92 dual sounding
Estimated uncertainty vs observed differences
Altitude [km]
Night (N=17)
Day (N=29)
Humidity
Re-assess RH
recalibration
Errors
• T-dependent
calibration
• Dry bias
• Time lag
Lindenberg
Ground check in SHC
• Traceability
• 4% change over ~9
years
• SHC readings enter
uncertainty budget
• Future: use SHC to
scale profile
Lindenberg
RH: Dry biases
 Heating of humidity sensor
- ΔT: radiation correction of T-sensor
- f: enhancement factor (laboratory experiments)
- Uncertainties: ΔT, f
 Temperature-dependent dry bias (-30oC to -70oC)
- Based on RS92 - CFH comparison
- Max at 7% at -60oC (similar to Voemel, 2007)
- Uncertainty: comparable to correction
RH: time-lag
Yangjiang 20 July 2010
Relevant below -40oC,
τ = 20s (τ > 100s @ -80oC)
Flattens features in humidity
profile
Correction:
numeric inversion of low-pass
filter. Enhances structures &
noise (a-posteriori filtering)
Uncertainties: time constant,
statistical noise
RH: corrections & uncertainties
Dominant uncertainties:
• Calibration
• Cal. correction
• Dry bias
Yangjiang 20 July 2010
RH: GRUAN – Vaisala
(Lindenberg)
• Calibration correction:
• Dry bias:
~5%
~10%
Altitude [km]
Night (N=277)
Day (N=258)
GRUAN achievements
• GRUAN data product for Vaisala RS92 radiosonde
• Other radiosonde products are developed (Modem M10, Meisei
RS11-G, Meteolabor SRS34, Frost point hygrometer)
• Other products & data streams under development :
• GNSS total water vapor column
• Raman Lidar (T, U)
• μ-wave radiometer (T, U)
• Archive with ~30,000 GRUAN-processed radiosounding profiles
with individual measurement uncertainties.
• > 20 GRUAN-related publications
GRUAN summary
 GRUAN is a new international approach to long term
observations of upper air essential climate variables
 Focus on priority 1 Essential Climate Variables
(ECVs) to start: Water vapour and temperature
 Goal is to provide reference observations:
- quantified uncertainties,
- traceable,
- well documented
 However, GRUAN receives no direct funding, so
relies on national / project support to realise its goals.
GAIA-CLIM Project
 NPL, UKMO and ECMWF are part of a
new three year Horizon2020 Earth
Observation project on ‘Gap Analysis
for Integrated Atmospheric ECV
CLImate Monitoring’ (GAIA-CLIM).
 The principal aim of GAIA-CLIM is to
“lead to a step change of availability of,
and ability to utilize, truly reference
quality traceable measurements in
support of satellite data
characterisation. It is only if robust
uncertainty estimates are placed on the
ground-based and sub-orbital data and
used in the analysis that unambiguous
interpretation of EO sensor
performance can occur.”
Instrument development at NPL
 The accurate determination of atmospheric temperature
and humidity is still a challenging measurement issue.
 This is particularly the case in the UT/LS where sensors
need fast response in low density air, and solar heating
and water contamination present additional problems for
both temperature and humidity sensors.
 Non-contact measurement methods offer the potential to
address these challenges.
 NPL are working on new rapid measurement
technologies.
 Includes work to integrate temperature and water vapour
analysis and assess overall uncertainty of combined
measurement.
Instrument design
• Non-Contact Thermometer and Hygrometer
(NCTAH)
• Acoustic temperature and laser absorption
humidity measurements in same air volume
• Open framework to minimise sampling effects
Questions ?