Progress on Nowcasting convection occurrence from space-borne - - PowerPoint PPT Presentation

Workshop on NWC application using MTG-IRS Darmstadt 25-26 July 2013

P . Antonelli(1) , A. Manzato(2) , T. Cherubini (3) ,

• S. Tjemkes (4), R. Stuhlmann(4) , E. Holm (5) ,
• C. Serio (6), G. Masiello (6)

(1) Space Science Engineering Center - University of Wisconsin - Madison (2) OSMER Arpa Friuli Venezia Giulia (3) Mauna Kea Weather Center (4) EUMETSAT (5) ECMWF (6) University of Basilicata

Progress on Nowcasting convection occurrence from space-borne instability predictors

Thursday, July 25, 13

History

Project started in 2010 and had contributions in various forms from several people from ARPA-FVG (OSMER), CNMCA, ITALIAN CIVIL PROTECTION, and SPACE SCIENCE ENGINEERING CENTER; Project aimed to derive instability indices from IASI Level 2 products, and to improve short term forecast of severe convective events; In 3 years project expanded to involve ECMWF (through EUMETSAT), AER (provided OSS), and University of Hawaii (MKWC); Project represents a significant effort in achieving standard needed to make products from high-spectral resolution IR data available to forecaster.

PART I

Previous Studes

Thursday, July 25, 13

Area of interest

Generation of IASI Full dataset PART I

Previous Studes

Thursday, July 25, 13

Strategy

• define occurrence of convective event, by setting threshold of 10 strikes to convert

discrete distribution of lightning strikes into binary output event yes/no (1/0);

• build Full Dataset with occurrence of convective event (yes/no) and values of all available

predictors;

• divide the full dataset into 2 subsets: 1) Total Set, to be used to build classifier; 2) Test Set, to be used for final evaluation of classifier

capacity of prediction;

• divide, in 12 different ways, Total Set into: 1) Training Set (75%); 2) Validation Set (25%); both to be used to sub-select the optimal

predictors using Repeated Holdout Technique [Witten:2005] PART I

Previous Studes

Thursday, July 25, 13

• divide inputs in 21 bins and for each bin calculate ratio:

Preprocessing ψi(x) = N act

i

N tot

i

PART I

Previous Studes

Thursday, July 25, 13

• divide inputs in 21 bins and for each bin calculate ratio:

Preprocessing

• these values are then fit with ad-hoc functions to define, for

each predictor (for example KI), empirical posterior probability (EPP(KI)), i.e. mathematical relationship which associate the probability of event=1 to continuous values of the predictors and use it as pre-processing;

ψi(x) = N act

i

N tot

i

PART I

Previous Studes

Thursday, July 25, 13

• divide inputs in 21 bins and for each bin calculate ratio:

Preprocessing EPP(KI) = (0.33 ∗ exp(0.074 ∗ (KI − 20)))− > x

• these values are then fit with ad-hoc functions to define, for

each predictor (for example KI), empirical posterior probability (EPP(KI)), i.e. mathematical relationship which associate the probability of event=1 to continuous values of the predictors and use it as pre-processing;

ψi(x) = N act

i

N tot

i

PART I

Previous Studes

Thursday, July 25, 13

• implement forward selection algorithm (based on Artificial Neural Networks, namely single

layer, feedforward network trained with backpropagation [Manzato:2004, Manzato-2007]) to choose optimal subset of predictors;

• ANN chooses at first one predictor that gives the best classification of the event
• ccurrences starting from its empirical probability distribution. Then it selects predictor

which gives best fit, when used together with first one. New predictors are added, until the system predictive skill stop increasing. During input selection process, number of hidden neurons varies according to predefined function of number of inputs;

• number of input predictors was chosen taking into consideration the mean skill of the 12

ANN built with the different instances of the Total Sets. Prediction skill of ANN was measured by the mean cross-entropy error (CEE):

• where is the output of the ANN, and is boolean for the the convective event (1|yes

0|no), calculated over the 12 instances of the Validation Sets;

Strategy CEE = − PN

n=1 [tnln (yn) + (1 − tn) ln (1 − yn)]

yn tn

PART I

Previous Studes

Thursday, July 25, 13

TV diagram

PART I

Previous Studes

Thursday, July 25, 13

TV diagram

PART I

Previous Studes

Thursday, July 25, 13

Strategy

• once optimal subset of predictors was identified, final ANN architecture was chosen

among different candidates (different numbers of hidden neurons in hidden layer) as one with lowest combined CEE on Total set (Training + Validation) without overfitting it, that is, with similar performances also on independent Test set;

H1 H1 Inputs

W11 X1 W62 X6 W11 W21

φ(Σwx+α) φ(Σwx+α) φ(Σwx+α)

• PART I

Previous Studes

Thursday, July 25, 13

Strategy

• quantitative evaluation of learning and generalization of knowledge during ANN supervised

training was performed using Relative Operating Characteristic (ROC) [Swets:1973]. Once

• utput of ANN was dichotomized using event prior probability as threshold, contingency

table was calculated, and different statistical scores were determined [Manzato: 2007].

Event (Y) Event (N) Prediction: (Y)

a b

Prediction: (N)

c d

PSS =

(ad-bc) (a+c)(b+d)

POD =

a a+c

POFD =

b b+d

FAR =

b a+b

PART I

Previous Studes

Thursday, July 25, 13

Event Climatology

PART I

Previous Studes

Thursday, July 25, 13

Results using RAWINSONDES

• During input selection phase (forward selection algorithm) only Total set was used, it

included 949 cases and was used to train different ANN candidates. While to select best architecture (hidden neurons) for the prediction system (ANN) also consistency between the results obtained on the Total and on the Test sets (of 350 cases) was taken into

• account. The architecture chosen was with 8 inputs, 2 neurons on the hidden layer, and 1
• utput.

TRAINING: Application of the ANN on the Total set led to a Total CEE of 0.335, while applying the probability threshold (0.40) on the continuous ANN output led to the following contingency table:

TOTAL Event (Y) Event (N) Prediction: YES 316 95 Prediction: NO 63 475

TOTAL POD HIT FAR POFD PSS Score 0.83 0.83 0.23 0.17 0.67

TESTING: Applying ANN on Test set led to a Test CEE of 0.375, while applying the probability threshold (0.40) on the continuous ANN output led to the following contingency table:

TEST Event (Y) Event (N) Prediction: YES 114 33 Prediction: NO 23 180 TEST POD HIT FAR POFD PSS Score 0.83 0.84 0.22 0.15 0.68

PART I

Previous Studes

Thursday, July 25, 13

Results using IASI data

• By focusing on single area of interest, it was possible to generate ANN trained on a IASI

dataset twice as large as the Full IASI Set. In this case the event occurrence was defined by at least 3 (IC+C2G) lightnings. Best ANN-inputs were ShowI, PCS MW 7, MRH. The best ANN, a 3 input, 1 hidden neuron.

Analysis of IASI results

TRAINING: Application of the ANN on the Total set led to a Total CEE of 0.30, on 1338 cases, while applying the probability threshold (0.14) on the continuous ANN output led to the following contingency table:

TOTAL Event (Y) Event (N) Prediction: YES 138 330 Prediction: NO 41 829

TOTAL POD HIT FAR POFD PSS Score 0.77 0.72 0.70 0.28 0.48

TESTING: Applying ANN on Test set led to a Test CEE of 0.36, on 657 cases, while applying the probability threshold (0.14) on the continuous ANN output led to the following contingency table:

TEST Event (Y) Event (N) Prediction: YES 92 197 Prediction: NO 17 349 TEST POD HIT FAR POFD PSS Score 0.84 0.67 0.68 0.36 0.48

PART I

Previous Studies

Thursday, July 25, 13

Outcome

• False alarm was found to be too high, and PSS was found to be too low;
• Results clearly indicated need for retrieval improvements;
• in terms of number of successful retrievals;
• in terms of retrieval accuracy;
• Areas of potential improvements identified:
• a-priori;
• surface emissivity representation;
• numerical stability;

PART I

Previous Studes

Thursday, July 25, 13

Retrieval of high spectral resolution observations over Hawai’i using WRF and ECMWF derived a-priori

Tiziana Cherubini (1) Paolo Antonelli (2)

(1) Muna Kea Weather Service - Univeristy of Hawai’i (2) Space Science Engineering Center - University of Wisconsin - Madison

Presented at MIST VIII - KNMI - 6-7 Dec 2012

Improvements in characterization of a-priori information

PART II

A-priori

Thursday, July 25, 13

Study over Hawai’i

• the main idea was to transition from a retrieval system based on climatological a-priori to a

system based on a NWP derived a-priori;

• transition steps:
• application of UWPHYSRET to a selected Hawai’i case (21 Jul 2012) using climatological a-priori derived from local rawinsonde

(same approached used over Udine), with retrievals being performed on IASI, CrIS, and AIRS data;

• application of UWPHYSRET to the same case using a single profile generated by the WRF model as First Guess and the simple

a-priori derived from the MKWC WRF (12hr forecast - analysis); Results presented at MIST VIII by Dr. Cherubini PART II

A-priori

Thursday, July 25, 13

CrIS Retrieval from Climatology

Temperature

PART II

A-priori

Thursday, July 25, 13

CrIS Retrieval from Climatology

Temperature

PART II

A-priori

Thursday, July 25, 13

CrIS Retrieval from WRF: a-priori

T WV log(q)

WV 12hr Forecast - Analysis T 12hr Forecast - Analysis

PART II

A-priori

Thursday, July 25, 13

IASI Retrievals from ECMWF (EF) a-priori

Temperature

PART II

A-priori

Thursday, July 25, 13

IASI Retrievals from ECMWF (EF) a-priori

WV

PART II

A-priori

Thursday, July 25, 13

IASI Retrievals from ECMWF a-priori

Over Land

Courtesy of S. Tjemkes

PART II

A-priori

Thursday, July 25, 13

Conclusions

• In order to improve accuracy of the retrievals UWPHYSRET moved from climatological a-

priori derived from local rawinsonde to NWP derived a-priori (work still in progress);

• first tests with model a-priori were done in collaboration with MKWC over Hawai’i using

CrIS, AIRS, and IASI data;

• use of NWP model apriori and first guess showed potential for significant improvement in

retrieval accuracy however more work is needed in optimizing the use of the model a-priori (which might exhibit extremely small values in terms of the variance of some of the atmospheric state variables);

• large differences between FG and final retrievals were found over land;

PART II

A-priori

Thursday, July 25, 13

Retrieval of high spectral resolution observations over Udine using ECMWF derived a-priori with improved Surface Emissivity characterization

Thursday, July 25, 13

Changes in UWPHYSRET

• smoothing and Inflating of ECMWF a-priori covariance;
• improvements in the calculation of Sa Inverse;
• changes in representation of Surface Emissivity elements in state vector:
• a-priori of Surface Emissivity derived from Dan Zhou’s atlas FOV by FOV;
• a-priori covariance of Surface Emissivity inflated;
• Observation error was obtained either inflating diagonal noise covariance by 30% or by

using full noise covariance from CNES;

PART II

A-priori

✏ → ln ✓ ✏ 1 − ✏ ◆ Changes have been tested on a variety of cases, this presentation focuses on Udine data which will be used for instability study

Thursday, July 25, 13

Modification to ECMWF apriori

PART II

A-priori

Thursday, July 25, 13

Modification to SE representation

PART II

A-priori

Thursday, July 25, 13

Case studies

• 16 June 2012
• overpass at 09:04:58 UTC
• 189 FOVs (140 clear according to AVHRR)
• 182 FOVs detected as clear sky from UWPHYSRET
• apriori from ECMWF high resolution forecasting system T1279 (SES)
• 01 April 2009
• overpass at 09:54:25 UTC
• 107 FOVs (all clear according to AVHRR)
• 33 FOVs detected as clear sky from UWPHYSRET
• apriori from ECMWF high resolution forecasting system T799 (deterministic), successfully

implemented on Wednesday, 1 February 2006. (EF)

PART III

Case Studies

Thursday, July 25, 13

16 June 2012: MODIS view

• 16 June 2012
• overpass at 09:04:58 UTC
• 189 FOVs (140 clear according to AVHRR)
• 182 FOVs detected as clear sky from UWPHYSRET
• apriori from ECMWF high resolution forecasting system

T1279 (SES) MODIS overpass at 09:55 UTC

PART III

Case Studies

Thursday, July 25, 13

16 June 2012: IASI FOVs

Retrieval Clear Retrieval Cloudy AVHRR Clear 140 AVHRR Cloudy 42 7 AVHRR CM gives false cloudy over coastal areas

PART III

Case Studies

Thursday, July 25, 13

16 June 2012: IASI individual RH retrieval

Closest FOV to RAOB site WET LAYER NOT RETRIEVED

PART III

Case Studies

Thursday, July 25, 13

16 June 2012: IASI individual T retrieval

Closest FOV to RAOB site

PART III

Case Studies

Thursday, July 25, 13

16 June 2012: IASI individual WV retrieval

Closest FOV to RAOB site WET LAYER NOT RETRIEVED

PART III

Case Studies

Thursday, July 25, 13

16 June 2012: IASI individual SE, SKT retrieval

Closest FOV to RAOB site

Thursday, July 25, 13

16 June 2012: RAOB

16044 LIPD Udine Observations at 12Z 16 Jun 2012

• PRES HGHT TEMP DWPT RELH MIXR DRCT SKNT THTA THTE THTV

hPa m C C % g/kg deg knot K K K

• 1010.0 94 26.8 14.8 48 10.58 180 2 299.1 330.3 301.0

1000.0 178 24.8 13.8 50 10.01 185 3 297.9 327.4 299.7 998.0 196 24.6 14.6 54 10.57 184 3 297.9 328.9 299.8 939.0 726 19.6 13.6 68 10.53 169 2 298.1 328.9 299.9 925.0 856 18.8 12.8 68 10.14 165 2 298.5 328.4 300.4 923.0 875 18.4 12.4 68 9.89 167 2 298.3 327.4 300.1 866.0 1419 15.4 7.4 59 7.51 212 4 300.7 323.2 302.0 850.0 1577 15.4 -0.6 33 4.33 225 4 302.3 315.7 303.1 841.0 1667 15.0 -4.0 27 3.40 225 4 302.8 313.4 303.4 832.0 1758 14.4 -3.6 29 3.54 225 5 303.1 314.2 303.7 823.0 1850 14.6 -5.4 25 3.12 225 5 304.2 314.1 304.8 743.0 2704 9.2 -14.8 17 1.64 225 9 307.4 312.8 307.7 729.0 2860 8.1 -13.7 20 1.84 225 10 307.9 313.9 308.2 700.0 3194 5.8 -11.2 28 2.33 245 15 308.9 316.5 309.3 637.0 3956 1.0 -14.7 30 1.93 255 18 311.8 318.3 312.2 607.0 4346 -1.5 -16.5 31 1.74 245 17 313.3 319.2 313.6 587.0 4612 -3.3 -24.3 18 0.92 239 16 314.2 317.5 314.4 576.0 4761 -3.7 -31.3 10 0.49 235 16 315.5 317.3 315.6 548.0 5154 -4.7 -49.7 2 0.07 244 14 318.8 319.1 318.8 545.0 5197 -5.0 -50.1 1 0.07 245 14 319.0 319.3 319.0 500.0 5870 -8.9 -55.9 1 0.04 235 19 322.1 322.3 322.1 476.0 6248 -11.1 -56.1 1 0.04 234 20 324.0 324.1 324.0 406.0 7439 -21.2 -45.1 10 0.17 230 23 326.0 326.7 326.1 400.0 7550 -22.1 -44.1 12 0.19 230 22 326.2 326.9 326.2 WET LAYER Wind direction Wind speed PART III

Case Studies

Thursday, July 25, 13

16 June 2012: IASI FOVs

Raob launch: 11:00 UTC Assuming:

• 1hr to reach 600 hPA
• wind speed of 18 knots
• wind dir of 245 degrees

Then:

• closest FOV (projected in

time and space) is FOV #136 at about 90 km from RAOB site PART III

Case Studies

Thursday, July 25, 13

16 June 2012: IASI individual RH retrieval

FOV 136 WET LAYER PROPERLY RETRIEVED

PART III

Case Studies

Thursday, July 25, 13

16 June 2012: IASI individual T retrieval

FOV 136

PART III

Case Studies

Thursday, July 25, 13

16 June 2012: IASI individual WV retrieval

FOV 136 WET LAYER PROPERLY RETRIEVED

PART III

Case Studies

Thursday, July 25, 13

16 June 2012: RETRIEVAL RH departure from FG

PART III

Case Studies

Thursday, July 25, 13

16 June 2012: RETRIEVAL T departure from FG

PART III

Case Studies

Thursday, July 25, 13

16 June 2012: RETRIEVAL WV departure from FG

PART III

Case Studies

Thursday, July 25, 13

16 June 2012: RETRIEVAL residuals stats

PART III

Case Studies

Thursday, July 25, 13

01 April 2009: MODIS view

MODIS overpass at 10:20 UTC

• 01 April 2009
• overpass at 09:54:25 UTC
• 107 FOVs (all clear according to AVHRR)
• 33 FOVs detected as clear sky from UWPHYSRET
• apriori from ECMWF high resolution forecasting

system T799 (deterministic), successfully implemented

• n Wednesday, 1 February 2006. (EF)

PART III

Case Studies

Thursday, July 25, 13

01 April 2009: IASI FOVs

Retrieval Clear Retrieval Cloudy AVHRR Clear 33 74 AVHRR CM = 0% for every FOV clearly is missing for 2009

PART III

Case Studies

Thursday, July 25, 13

01 April 2009: IASI individual RH retrieval

Closest FOV to RAOB site

PART III

Case Studies

Thursday, July 25, 13

01 April 2009: IASI individual T retrieval

Closest FOV to RAOB site

PART III

Case Studies

Thursday, July 25, 13

01 April 2009: IASI individual WV retrieval

Closest FOV to RAOB site

PART III

Case Studies

Thursday, July 25, 13

01 April 2009: IASI individual SE, SKT retrieval

Closest FOV to RAOB site

PART III

Case Studies

Thursday, July 25, 13

01 April 2009: IASI individual spectral residual for retrieval

Closest FOV to RAOB site

PART III

Case Studies

Thursday, July 25, 13

01 April 2009: RETRIEVAL RH departure from FG

PART III

Case Studies

Thursday, July 25, 13

01 April 2009: RETRIEVAL T departure from FG

PART III

Case Studies

Thursday, July 25, 13

01 April 2009: RETRIEVAL WV departure from FG

PART III

Case Studies

Thursday, July 25, 13

01 April 2009: RETRIEVAL residuals stats

PART III

Case Studies

Thursday, July 25, 13

Issues with cloudy scene on 01 April 2009

Thursday, July 25, 13

01 April 2009: IASI individual T retrieval

Closest FOV to RAOB site Oscillation not explained

PART III

Case Studies

Thursday, July 25, 13

01 April 2009: IASI individual T retrieval

Closest FOV to RAOB site Retrieval maintain same vertical structure of FG, simply shifts profile towards warmer side

PART III

Case Studies

Thursday, July 25, 13

01 April 2009: IASI individual RH retrieval

FOV 94 FOV 94, from MODIS image, is expected to be completely

• vercast, however RH<100%

PART III

Case Studies

Thursday, July 25, 13

01 April 2009: IASI individual Τ retrieval

FOV 94 FOV 94, from MODIS image, is expected to be completely

• vercast, however T profile seems

to be realistic and close to FG (Δ<3σ)

PART III

Case Studies

Thursday, July 25, 13

01 April 2009: IASI individual WV retrieval

FOV 94 FOV 94, from MODIS image, is expected to be completely

• vercast, however WV profile

seems to be realistic and close to FG (Δ<3σ)

PART III

Case Studies

Thursday, July 25, 13

01 April 2009: IASI individual SE, SKT retrieval

FOV 94 FOV 94, from MODIS image, is expected to be completely

• vercast. SKT and SE seem to

provide only hints that FOV is clouldy: (Δskt>3σ), and SE unrealistic.

PART III

Case Studies

Thursday, July 25, 13

Preliminary conclusions

This work indicated that:

• 1. UWPHYSRET seems performing correctly over clear sky;
• 2. Convergence rate, Saturation, T and WV deviation from FG within 3σ, spectral residuals within

noise, are not sufficient to screen out cloudy cases. SE and SKT might provide further useful information to discriminate cloudy retrievals from clear ones;

• 3. open issues:
• high tropospheric oscillation in T profile for one of the cases could not be explained;
• vertical structure from FG is generally maintained by retrieval, even if not correct;
• cloudy FOVs contaminated statistics for 01 April 2009 case.
• 4. besides previous point, deviation from FG does not show pathological behavior;

PART III

Case Studies

Thursday, July 25, 13

Monthly statistics for retrieval deviation from FG

Thursday, July 25, 13

Monthly statistics: deviation from FG

2008 04

Text

PART IV

Retrieval Statistics

Thursday, July 25, 13

Monthly statistics: deviation from FG

2008 04 2008 05

Text

PART IV

Retrieval Statistics

Thursday, July 25, 13

Monthly statistics: deviation from FG

2008 04 2008 05 2008 06

Text

PART IV

Retrieval Statistics

Thursday, July 25, 13

Monthly statistics: deviation from FG

2008 04 2008 05 2008 06 2008 07

Text

PART IV

Retrieval Statistics

Thursday, July 25, 13

Monthly statistics: deviation from FG

2008 04 2008 05 2008 06 2008 07

Text

2008 08

PART IV

Retrieval Statistics

Thursday, July 25, 13

Monthly statistics: deviation from FG

2008 04 2008 05 2008 06 2008 07

Text

2008 08 2009 04

PART IV

Retrieval Statistics

Thursday, July 25, 13

Monthly statistics: deviation from FG

2008 04 2008 05 2008 06 2008 07

Text

2008 08 2009 04 2009 05

PART IV

Retrieval Statistics

Thursday, July 25, 13

Monthly statistics: deviation from FG

2008 04 2008 05 2008 06 2008 07

Text

2008 08 2009 04 2009 05 2009 06

PART IV

Retrieval Statistics

Thursday, July 25, 13

Monthly statistics: deviation from FG

2008 04 2008 05 2008 06 2008 07

Text

2008 08 2009 04 2009 05 2009 06 2009 07

PART IV

Retrieval Statistics

Thursday, July 25, 13

Monthly statistics: deviation from FG

2008 04 2008 05 2008 06 2008 07

Text

2008 08 2009 04 2009 05 2009 06 2009 07 2009 08

PART IV

Retrieval Statistics

Thursday, July 25, 13

Monthly statistics: deviation from FG

2008 04 2008 05 2008 06 2008 07

Text

2008 08 2009 04 2009 05 2009 06 2009 07 2009 08 2009 09

PART IV

Retrieval Statistics

Thursday, July 25, 13

Monthly statistics: deviation from FG

2008 04 2008 05 2008 06 2008 07

Text

2008 08 2009 04 2009 05 2009 06 2009 07 2009 08 2009 09 2011 04

PART IV

Retrieval Statistics

Thursday, July 25, 13

Monthly statistics: deviation from FG

2008 04 2008 05 2008 06 2008 07

Text

2008 08 2009 04 2009 05 2009 06 2009 07 2009 08 2009 09 2011 04 2011 05

PART IV

Retrieval Statistics

Thursday, July 25, 13

Monthly statistics: deviation from FG

2008 04 2008 05 2008 06 2008 07

Text

2008 08 2009 04 2009 05 2009 06 2009 07 2009 08 2009 09 2011 04 2011 05 2011 06

PART IV

Retrieval Statistics

Thursday, July 25, 13

Monthly statistics: deviation from FG

2008 04 2008 05 2008 06 2008 07

Text

2008 08 2009 04 2009 05 2009 06 2009 07 2009 08 2009 09 2011 04 2011 05 2011 06 2011 07

PART IV

Retrieval Statistics

Thursday, July 25, 13

Monthly statistics: deviation from FG

2008 04 2008 05 2008 06 2008 07

Text

2008 08 2009 04 2009 05 2009 06 2009 07 2009 08 2009 09 2011 04 2011 05 2011 06 2011 07 2011 08

PART IV

Retrieval Statistics

Thursday, July 25, 13

Monthly statistics: deviation from FG

2008 04 2008 05 2008 06 2008 07

Text

2008 08 2009 04 2009 05 2009 06 2009 07 2009 08 2009 09 2011 04 2011 05 2011 06 2011 07 2011 08 2011 09

PART IV

Retrieval Statistics

Thursday, July 25, 13

Time series of Instability Indices

Thursday, July 25, 13

Stability Indices

PWE from IASI L2 PWE from ECMWF FG

2011 August 28 @ 08:24:50 UTC

PART IV

Retrieval Statistics

Thursday, July 25, 13

Time series of Instability Indices

PART IV

Retrieval Statistics

Thursday, July 25, 13

Time series of Instability Indices

MRH LRH PART IV

Retrieval Statistics

Thursday, July 25, 13

Linear correlation between Stability Indices

04 2008 05 2008 06 2008 07 2008 08 2008 04 2009 05 2009 06 2009 07 2009 08 2009 09 2009 04 2011 05 2011 06 2011 07 2011 08 2011 CAPE 0.77 0.12 0.25 0.69 0.53 0.47 0.63 0.51 0.32 0.61 0.00 0.17 0.48 0.94 0.30 0.29 0.23 0.31 0.37 0.55 0.26 0.22 0.41 0.06 0.11 0.75 0.52 0.55 0.53 0.77 0.50 0.59 CIN 0.02 0.08 0.10 0.13 0.09 0.05 0.64 0.06 0.21 0.08 0.16 0.35 0.24 0.26 0.36 0.40 0.67 0.68 0.32 0.48 0.12 0.30 0.61 0.39 0.32 0.44 0.67 0.66 0.43 0.24 0.80 0.89 DT500 0.13 0.28 0.22 0.52 0.66 0.67 0.58 0.49 0.42 0.56 0.33 0.58 0.38 0.74 0.41 0.44 0.56 0.66 0.56 0.73 0.42 0.56 0.79 0.86 0.76 0.82 0.89 0.94 0.65 0.82 0.81 0.86 DTC 0.09 0.10 0.22 0.43 0.62 0.59 0.63 0.49 0.47 0.53 0.30 0.43 0.42 0.62 0.33 0.37 0.61 0.68 0.62 0.77 0.67 0.57 0.90 0.88 0.69 0.85 0.88 0.92 0.69 0.74 0.80 0.85 KI 0.67 0.25 0.66 0.36 0.59 0.00 0.76 0.68 0.58 0.44 0.62 0.53 0.20 0.26 0.59 0.49 0.59 0.53 0.61 0.40 0.67 0.55 0.84 0.72 0.71 0.72 0.70 0.68 0.40 0.37 0.84 0.81 LI 0.27 0.23 0.26 0.61 0.72 0.66 0.45 0.38 0.42 0.49 0.24 0.52 0.35 0.73 0.61 0.46 0.54 0.68 0.59 0.76 0.34 0.58 0.83 0.90 0.79 0.85 0.83 0.86 0.60 0.82 0.79 0.83 LRH 0.51 0.48 0.71 0.63 0.72 0.33 0.69 0.21 0.36 0.09 0.54 0.59 0.04 0.30 0.29 0.38 0.47 0.30 0.41 0.27 0.74 0.47 0.78 0.77 0.45 0.48 0.43 0.35 0.49 0.60 0.80 0.77 MRH 0.66 0.21 0.77 0.58 0.78 0.27 0.56 0.01 0.65 0.24 0.76 0.49 0.13 0.01 0.76 0.55 0.54 0.28 0.65 0.26 0.86 0.57 0.82 0.78 0.75 0.66 0.80 0.68 0.82 0.67 0.93 0.88 PWE 0.50 0.25 0.61 0.58 0.57 0.39 0.76 0.51 0.74 0.52 0.34 0.20 0.70 0.77 0.72 0.71 0.61 0.52 0.51 0.28 0.78 0.64 0.74 0.74 0.81 0.78 0.48 0.40 0.81 0.80 0.86 0.85 ShowI 0.18 0.23 0.42 0.37 0.63 0.44 0.70 0.34 0.59 0.35 0.33 0.36 0.56 0.64 0.24 0.54 0.64 0.78 0.53 0.44 0.36 0.63 0.87 0.81 0.82 0.90 0.71 0.80 0.55 0.79 0.67 0.73 THETAE 0.70 0.59 0.77 0.91 0.81 0.78 0.69 0.55 0.57 0.67 0.24 0.57 0.80 0.94 0.68 0.76 0.70 0.67 0.60 0.72 0.69 0.78 0.81 0.94 0.92 0.95 0.84 0.83 0.81 0.95 0.89 0.92 T799 T1279 PART IV

Retrieval Statistics

Thursday, July 25, 13

Conclusions

• Inversion of high spectral resolution data is constantly improving;
• There is still space for significant improvement of characterization of a-priori covariances;
• The most significant source of noise, in terms of L2 products, is related to uncertainties on

surface emissivity and cloud contamination;

• L2 data generated from high spectral resolution infrared radiances are capable of

characterizing pre-convective environment, especially in terms of water vapor;

• The use of L2 by the nowcasting community is the key in achieving the accuracy needed

to make the product useful.

Thursday, July 25, 13

Some references

Agostino Manzato, G. Morgan Jr. 2003: Evaluating the sounding instability with the Lifted Parcel Theory. Atmospheric Research 67–68 (2003) 455–473. Agostino Manzato, 2001: A climatology of instability indices derived from Friuli Venezia Giulia soundings, using three different methods Atmospheric Research 67–68 (2003) 417– 454 Agostino Manzato, 2001: A Verification of Numerical Model Forecasts for Sounding- Derived Indices above Udine, Northeast Italy Weather Forecasting (2007) 477–495

• S. Puca, F. Zauli, L. De Leonibus, P

. Rosci, and P . Antonelli 2009. Automatic detection and monitoring of convective cloud systems based on geostationary infrared observations. Submitted to Meteorological Applications, 2009. Silvia Puca, Daniele Biron, Luigi De Leonibus, Paolo Rosci, and Francesco Zauli 2005. Improvements on numerical "objects" detection and nowcasting of convective cell with the use of seviri data (ir and wv channels) and numerical techniques. In The International Symposium on Nowcasting and Very Short Range Forecasting (WSN05), 2005.C. Rodgers 2000: Inverse Methods for Atmospheric Soundings. World Scientific.

Thursday, July 25, 13

Some references

David Tobin, Paolo Antonelli, Henry Revercomb, Steven Dutcher, David Turner, Joe Taylor, Robert Knuteson, and Kenneth Vinson, 2007: Hyperspectral Data Noise Characterization using Principle Component Analysis: Application to the Atmospheric Infrared Sounder. J. Appl. Remote Sens. 1, 013515 (2007) doi: 10.1117/1.2757707 Antonelli P ., H. E. Revercomb, W. L. Smith, R.O. Knuteson, L. Sromovsky, D.C. Tobin, R. K. garcia, H. B. Howell, H.-L. Huang, F.A. Best, 2004: A Principal Component Noise Filter for High Spectral Resolution Infrared Measurements. J. Geophys. Res.,109, D23102, doi: 10.1029/2004JD004862. Huang H.-L., P . Antonelli, 2001: Application of Principal Component Analysis to high resolution infrared measurements compression, and retrieval. J. Appl. Meteor., 40:25, pp. 365-388 Turner, D.D., R.O. Knuteson, H.E. Revercomb, C. Lo, and R.G. Dedecker, 2006: Noise reduction

• f Atmospheric Emitted Radiance Interferometer (AERI) observations using principal component
• analysis. J. Atmos. Oceanic Technol., 23, 1223-1238.

Swets, J. A., 1973: The relative operating characteristic in psychology. Science, 182, 900–1000.

Thursday, July 25, 13