On the Generation of an Optimized Fractional Cloudiness Time Series using a Multi-Sensor Approach

TECO-2010, Helsinki | 31 August 2010

On the Generation of an Optimized Fractional Cloudiness Time Series using a Multi-Sensor Approach Wiel Wauben*, Marijn de HaijReinout Boers, Henk Klein Baltink, Bert van Ulft,Mark Savenije

*R&D Information and Observation Technology, Climate Observations Dept, Regional Climate Dept, Weather Research and Development Dept


Contents

• Introduction

• Instruments

• Combination algorithm

• Cabauw Fractional Cloudiness

• Conclusions and outlook


3

Cabauw Experimental Site for Atmospheric Research

• Five remote sensing techniques for cloud observations

• Active and passive

• Column and hemispheric (integrated and resolved/scanning)

• 1 year data sets of 10-minute cloud data (15 May 2008 - 14 May 2009, total cloud cover & base)

• Generation of optimized & continuous cloudiness time series

• Evaluation of different techniques

CESAR


4

35 GHz cloud radar & CT75K

Instruments

Ceilometer (operational SYNOP/METAR cloud product)

• Sensitive to detect high cirrus• CLOUDNET procedure

• column techniques• including cloud base height


5

Instruments

•long-wave downward radiation•integrated hemispheric•APCADA algorithm

•thermal infrared•scanning•cloud mask

•visual digital camera•cloud mask•day-time only

Pyrgeometer (BSRN)

NubiScope

Total sky imager (TSI)


6

Manual approach

Combination

Goal: construction of optimized & continuous cloudiness time series

• strong / weak points• situation dependent• subjective• no reference!• complex algorithm• not generic

• “simple” weightedaverage based on experiences

• checked with climatologyof manual observations (1970-2000)

Hence


7

“Reference” algorithm

Combination

ijiji

ijijiji

j WH

CWHR

,,

,,,•

• Rj is the reference cloudiness (in percentage) at time j

• Wi,j is the weighting value at time j for the i-th instrument

• Hi,j=1 when the i-th instrument has a valid output at time j, else =0

• Ci,j is the cloudiness (in percentage) measured by the i-th instrument at time j

• WNUB,j = WTSI,j = 1

• for APCADA, CLOUDNET, LD40

• DCLOUDNET,j is the observed minimum CLOUDNET cloud base height in the 10-minute period at time j

• uncertainty for all

)3.1/exp( ,, kmDW jCLOUDNETji

2/12,

,

)(1

jiIj

ji CRN

E 0, jij CR


8

Cabauw fractional cloudiness

Cloud cover histogram

• Column

n=0, 8 high

n=2-7 low

• CLOUDNET

60% n=8

mainly due

to cirrus

• “reference” is

good

compromise

• low n=2-6 (higher during day time)


9


Cloudiness versus cloud base height

• NubiScope & TSI

generally best

agreement

• ACPADA &

LD40 lower

• CLOUDNET

too high


10


Contingency matrix LD40 () versus Reference ()

0 1 2 3 4 5 6 7 8 % #

0 10.5 6.1 1.2 0.8 0.4 0.3 0.4 0.2 0.1 20.1 10309

1 2.1 3.8 1.1 0.7 0.4 0.3 0.4 0.3 0.1 9.3 4769

2 0.3 0.8 0.8 0.5 0.3 0.2 0.2 0.3 0.0 3.3 1700

3 0.1 0.4 0.6 0.6 0.3 0.2 0.2 0.3 0.0 2.8 1445

4 0.1 0.2 0.3 0.8 0.6 0.4 0.4 0.5 0.0 3.3 1706

5 0.0 0.1 0.1 0.3 0.7 0.6 0.5 0.7 0.0 3.1 1574

6 0.0 0.1 0.1 0.1 0.4 0.7 0.7 1.2 0.0 3.3 1712

7 0.0 0.0 0.1 0.1 0.3 0.8 1.8 6.0 1.8 10.9 5575

8 0.0 0.0 0.0 0.1 0.1 0.3 0.9 10.6 32.0 44.0 22592

% 13.2 11.6 4.3 3.9 3.5 3.8 5.4 20.2 34.0 100.0

# 6777 5981 2201 2024 1812 1972 2790 10377 17448 51382

Band 0: 55.6% Band 1: 85.6% Band 2: 92.4% Over: 2.0% Under: 5.5% Δ: -0.1 |Δ|: 0.7

• 8 % with differences > 2 okta; fraction clear sky & overcast


11


Reference data set cloudiness

• 98% availability10-minute cloudiness

• e.g. daily with uncertainty


12

Conclusions

• Reference is weighted combination of individual instruments• Not a true reference, but general and robust approach that

produces useful results• Compromise whereby the NubiScope and TSI are considered to be

a higher quality product (weight 1) than the others (height dependent weight)

• Uncertainty range of reference cloudiness determined from the negative and positive differences between the reference and the cloudiness reported by each instrument over the time period under consideration

• Findings for instruments see paper• OBS also has limitations so 100% similarity not expected• Automated cloudiness using ceilometer introduced changes in

climatological cloud observations records

Conclusions & Outlook


13

Outlook

• APCADA and TSI are being / have been optimised as a result of this study

• Physical definition of cloud/cloudiness, threshold possibly dependent on application

• Usage of hemispheric method to overcome changes in climatological cloud observations records should be considered

• Towards scanning reference system?

Conclusions & Outlook


14

Thank you for your attention!

Lookup conference paper for more information

Boers, R., M.J. de Haij, W.M.F. Wauben, H. Klein Baltink, L.H. van Ulft, M. Savenije and C.N. Long (2010),

Optimized Fractional Cloudiness Determination from Five Ground - based Remote Sensing Techniques,

submitted to J. Geophys. Res.

On the Generation of an Optimized Fractional Cloudiness Time Series using a Multi-Sensor Approach

Documents

Transcript of On the Generation of an Optimized Fractional Cloudiness Time Series using a Multi-Sensor Approach