cln qa/qc efforts

18
CLN QA/QC efforts CCNY – (Barry Gross) UMBC- (Ray Hoff) Hampton U. (Pat McCormick) UPRM- (Hamed Parsiani)

Upload: paniz

Post on 01-Feb-2016

64 views

Category:

Documents


1 download

DESCRIPTION

CLN QA/QC efforts. CCNY – (Barry Gross) UMBC- (Ray Hoff) Hampton U. (Pat McCormick) UPRM- (Hamed Parsiani). Outline. “Raw” signal tests. Matchups against Rayleigh Linearity tests with ND filters Member processing algorithms Efforts to test algorithms for consistency - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: CLN QA/QC efforts

CLN QA/QC efforts

CCNY – (Barry Gross)UMBC- (Ray Hoff)

Hampton U. (Pat McCormick)UPRM- (Hamed Parsiani)

Page 2: CLN QA/QC efforts

Outline• “Raw” signal tests.

– Matchups against Rayleigh – Linearity tests with ND filters

• Member processing algorithms

• Efforts to test algorithms for consistency

• Indirect (Downstream) tests for retrieval accuracy

• Potential QA/QC efforts for CLN

Page 3: CLN QA/QC efforts

Testing multi-wavelength lidar signals to the molecular reference

0 2 4 6 8 10 12

10-2

100

102

Altitude (km)

Lida

r si

gnal

s (m

V)

CCNY-lidar, 20060329: 4:24pm (30 min. ave)

Lidar-355molecule

0 2 4 6 8 10 12

10-2

100

102

Altitude (km)

Lid

ar

sig

na

ls (

mV

)

CCNY-lidar, 20060329, 4:24pm (30 min. ave)

Lidar-532molecule

0 2 4 6 8 10

10-2

100

102

Altitude (km)

Lida

r si

gnal

s (m

V)

Lidara-1064molecule

CCNY-lidar, 20060329, 4:24pm (30 min. ave)

Lidar System Calibration Regression at 10-11 km

Representative matching of lidar profiles with Molecular profiles

Page 4: CLN QA/QC efforts

Good linearity!

NDF-1 (OD=1.6) at 12:56 pmNDF-2 (OD=1.0) at 12:59 pm

Lidar signal profiles

Lidar signal ratio

Lidar signal linearity: signal profiles and their ratios

Page 5: CLN QA/QC efforts

CCNY Processing

• Standard processing for 355 and 532 channels using Fernald Back-Integration method with S ratio pinned by AERONET AOD closure

• Far end Scattering Ratio Condition (1.01 at 355nm, 1.06 at 532 nm)

• Zmax determined by “minimum signal” method• 1064 channel uses system constant based on

cirrus cloud calibration

Page 6: CLN QA/QC efforts

CCNY Lidar Algorithm and Cross-Testing Efforts

• Different algorithms tested against each other.– Intercompare iterative and Fernald solutions

• Consistency check – Compare Measured Signal with Retrieved Signal after

optical property retrieval• 1064 channel system constant evaluation over

long time periods• Indirect assessment of standard Mie and

Raman optical properties using thin Cloud Optical Depth retrievals.

• Some preliminary cross-matchups with UPRM.

Page 7: CLN QA/QC efforts

Validation (1)Fernald vs Iterative

Range

Blue=exact FernaldGreen=iterative approximations

N=2

N=5

N=10

N=20

Page 8: CLN QA/QC efforts

Validation (2)Consistency Check Comparison of theoretical and

Measurement Signal

RP

RP

eqnLidar

meas

532nm

355 nmErrors < .3%

Page 9: CLN QA/QC efforts

Long term stability and evaluation of Lidar System Ratio

Page 10: CLN QA/QC efforts

Indirect Check of Optical Property retrieval using Cloud Optical Depth

• Raman COD retrieval based on successful derivation of cloud extinction and integrating

• Mie COD based on S. Young regression method and uses aerosol backscatter corrections above and below cloud

Clear sky for aerosol backscatter correction to COD

Page 11: CLN QA/QC efforts

Cross-Testing of Retrieval Algorithmson same Data

10-6

10-5

10-4

10-3

10-2

0

1

2

3

4

5

6

7

8

a [km-1

sr]

Altit

ude

(km

)

Aerosol Backscatter Coefficient

355

5321064

CCNY Processing UPRM Processing

Page 12: CLN QA/QC efforts
Page 13: CLN QA/QC efforts
Page 14: CLN QA/QC efforts
Page 15: CLN QA/QC efforts
Page 16: CLN QA/QC efforts
Page 17: CLN QA/QC efforts

Extra slides

Page 18: CLN QA/QC efforts

Test of lidar signal linearity at 355-nm

1. Time and date: 1256PM--1259PM, April 21, 2006 2. Method: Insert the different Neutral density Filters (NDF) in front of interference filter and PMT. Background level is calculated from the average of last 5-km lidar raw data. Mean and standard deviation are given. Signal ratios are calculated with the different NDFs. Their ratios should be the constant if both two signals are in the linear ranges. All data are the 2-min average lidar signal profiles. please note: ignore the variability of atmosphere and laser power.3. For the NDF, higher optical density (OD) values correspond to the

LOWER transmittances.