Accuracy assessment of an integrated profiling technique for temperature, humidity and liquid water content profiles

1
Accuracy assessment of an integrated profiling
technique for temperature, humidity and liquid
water content profiles
U. Lohnert (1), E. van Meijgaard (2), H. Klein
Baltink (2), S. Gross (1), R. Boers (2)
(1) Meteorological Institute, University of
Munich, Germany, (2) KNMI, De Bilt, Netherlands,
(boers_at_knmi.nl / Phone [phone])
2
Remote Sensing of atmospheric structure Why?

1. Possible replacement of radiosondes
2. Continuous monitoring of atmospheric stability
   indices
3. Continuous monitoring of radiative flux structure

3
Integrated Profiling Technique

• Recover the atmospheric state from remote sensing
  observations
• Combine remote sensing observations with a priori
  information (i.e. information about the
  atmospheric state prior to the measurements)
• No unique solution!

4
Remote Sensing of atmospheric structure The
instruments

1. Multiwavelength microwave radiometer (19 ch)
2. Cloud radar (35 GHz)
3. Cloud lidar
4. Boundary layer T, q (mast values)

5
Problem in testing IPT Atmospheric structure
can never be completely specified beforehand
gtgtSo, very difficult to evaluate with real data!

• Solution (not perfect)
• Impose the real world on the IPT-routine by means
  of a controlled model experiment
• Assure that the model represents natural
  variability as found in the real world

6
How good is the method to obtain atmospheric
thermodynamics and liquid water profiles?
Define truth (Regional Climate Model)
Forward modelling of microwave brightness
temperature
Simulating radar reflectivitites
Obtain representation of the truth
Backward modelling of T, q, q_l profiles (IPT)
7
Integrated Profiling Technique (IPTRL)
14 HATPRO brightness temperatures (TB)
dabs profiles

• optimized profiles of
• temperature (T)
• humidity (q)
• LWC (on variable radar resolution)

Radar-Lidar Ratio
Bayesian Retrieval
a priori LWC profile (mod. adiabatic)
a priori T und q profiles (nearest-by radiosonde)
measurement-consistent with respect to error
covariances
8
IPT-equations
y measurement vector (TB, dBZ, qcloud, Tgr,
qgr) Forward model F K Jacobi-matrix
(dF/dx) xa a priori profile Se error
covariance matrix Sa a priori covariance
IPT equations (optimal estimation) derivable
from Bayesian probability theory condition
Gaussian distributed parameters

• solution is not exact, but rather probability
  density
• a priori profile stabilizes the solution
• iterative approach guarantees physical
  consistency

9
LWC - a priori Sa
a priori profile modified adiabatic approach
(Karstens et al. 1994), based on empirical
aircraft measurements
LWC a priori covariance need to know how
accurate the a priori formulation is ... however,
no evaluation available instead vary a randomly
to a certain degree and then calculate
covariance
certain degree of ambiguity ... to be solved in
future
10
3 modes of application to liquid clouds

• Mode 0 no application possible due to
• Not all information available (e.g. missing
  measurements)
• Rain detected
• Melting layer within profile
• Mode 1 clear sky / pure ice cloud ? no cloud
  microphysics calculated, just temperature and
  humidity profile
• Mode 2 full IPT application ? use of a priori
  profile depends on the presence of small cloud
  droplets (SCD)
• if SCD present minimization in TB, dBZ, Tap,
  qap, LWCap
• if SCD not present ? minimization only in TB,
  dBZ, Tap, qap

11
Results

• Now-casting mode A priori data available at 0
  LT, 12 LT, prediction is done for times after the
  radiosonde launch
• Climate mode A priori is interpolated between 0,
  12 LT, reconstuction is performed
• First, let location of the a priori information
  be the same as the location where the IPT is
  tested
• Next, let location of the a priori information be
  different from the location where the IPT is
  tested

12
Nowcasting mode comparison with original model
data

• Blue is the a priori accuracy
• Red is IPT accuracy
• Magenta is statistical retrieval accuracy
• Green is theoretical accuracy
• Difference between Blue and Red is INFORMATION
  GAIN

13
Climate mode comparison with original model data

• Blue is the a priori accuracy
• Red is IPT accuracy
• Magenta is statistical retrieval accuracy
• Green is theoretical accuracy
• Difference between Blue and Red is INFORMATION
  GAIN

14
Climate mode comparison with original model
data (LWC)

1. Blue is the mean LWC profile
2. Red is IPT accuracy
3. Magenta is Z-profile is scaled with LWP
4. Green is theoretical accuracy

15
IPT accuracies as function of temporal-spatial
distance to radiosonde launch

• Temperature
• Blue is RMS assuming radiosonde persistent valid
• Red is RMS IPT error
• Dark green is Absolute Bias assuming radiosonde
  persistent valid
• Light green is Absolute Bias IPT

16
IPT accuracies as function of temporal-spatial
distance to radiosonde launch

• Absolute humidity
• Blue is RMS assuming radiosonde persistent valid
• Red is RMS IPT error
• Dark green is Absolute Bias assuming radiosonde
  persistent valid
• Light green is Absolute Bias IPT

17
Conclusions (1)

• Significant information gain to profiles in
  between radiosonde launches
• Information gain for T, q is larger in the
  Nowcasting mode than in the Climate mode
• IPT improves very much on a statistical retrieval
  technique for T, q
• Nowcasting application RMS 1.1K, 0.7 g m-3
• Climate application RMS 0.4K, 0.5 g m-3
• LWC accuracy improvement is not dependent on
  model of operation (lt 25 mg m-3)
• LWP accuracy below 10 gm-2

18
Conclusions (2)

• For absolute accuracies of T 1 K or less, then
  profiler should be located within 200 300 km of
  radiosonde launch site
• For absolute accuracies of q 0.5 g m-3, then
  profiler should be located within 100 km of
  radiosonde launch site
• A statistical retrieval technique for T yields
  the same accuracy as if the radiosonde was
  located within a distance of 300 km
• A statistical retrieval technique for q yields
  the same accuracy as if the radiosonde was
  located within a distance of 400 km

19
Outlook

• IPT can retrieve q, T, but also LWC, multiple
  clouds. Do cloud climatologies
• High resolution variability
• Add spatial-temporal information complement
  existing radiosonde launches
• May replace radiosonde launches (100 200 km)
• Can be expanded to include other types of
  information (radiative transfer applications), to
  be installed at Cabauw this month

20
Cloud radar data
21
Baseline Surface Radiation Network