Cloud determonations from surface solar radiation measruements

1
Cloud determonations from surface solar radiation
measruements

• Giles Harrison, Nicky Chalmers, Robin Hogan

2
Motivation

• Provide cloud observations at un-manned stations
• Make cloud observations more quantitative
• Use previous time-series of surface solar
  radiation to deduce information of how cloud
  cover has changed in the recent past
• Data required has been collected consistently in
  the UK since 1947

3
Pyranometers

• Sources of error
• Shade ring
• Freezing/thawing
• Buildings / trees

Plastic dome transmits 0.3-3µm
4
Typical clear conditions

• Diffuse fraction (Sd /Sg)
• Low in clear skies (approximately 0.2)
• Diurnal variation caused by changing solar zenith
  angle
• Opaqueness (1 G/E)
• Shows similar behavior to DF
• Always larger than DF in clear skies

5
Typical overcast conditions

• Diffuse fraction (Sd /Sg)
• Approaches 1 in cloudy skies
• Small variability
• Opaqueness (1 G/E)
• Always smaller than DF in clear skies
• Larger variability than DF proportional to
  cloud thickness?

6
Typical partially cloudy conditions

• Diffuse fraction (Sd /Sg)
• Large variability
• Mean value proportional to cloud amount?
• Opaqueness (1 G/E)
• Shows similar behavior to DF

7
Why use the diffuse fraction?
8
Cloud Amount

• Diffuse fraction is most sensitive to cloud
  amount between 2 and 6 oktas
• Outside of this region there are clear and
  overcast limits
• Opaqueness is most sensitive to cloud amount from
  5 oktas upwards
• It is more sensitive to cloud thickness
• Variability is important

9
Cloud Type

• Points are median values, bars are 5 confidence
  intervals for each category
• Cloud amounts are distinguishable to within human
  error bars
• Stratiform and cumulus cloud can be distinguished
• Stratus and Stratoculumuls cant be
  distinguished.

10
Conclusions

• The variability of the diffuse fraction and
  opaqueness as well as the average value is
  important for cloud determination
• Cloud determinations cannot be directly compared
  to human observations
• However they can be used as an independent,
  quantitative measure of cloud amount as in
  Harrison and Stephenson 2005
• Cloud form can be determined using the diffuse
  fraction
• The opaqueness can be used as a measure of cloud
  thickness