CLOUD IDENTIFICATION USING SURFACE SOLAR RADIATION MEASUREMENTS

1

CLOUD IDENTIFICATION USING SURFACE SOLAR
RADIATION MEASUREMENTS
Nicky Chalmers Supervisors Dr. Giles Harrison
and Dr. Robin Hogan
Two Kipp and Zonen CM5 pyranometers with a shade
ring mounted on the one in the foreground to
obtain the diffuse radiation. Accuracy 3 (Kipp
Zonen 2000)
Luke Howard Born in London 1772, invented the
naming convention used to classify clouds which
is still used today (Hamblyn, 2001).
Abstract Surface solar radiation was shown to
have the potential to be used as a measurement of
cloud amount and type. The total downwelling
surface solar radiation (global radiation) was
measured using a pyranometer. The direct
component (radiation received directly from the
sun) was removed using a second pyranometer with
a shade ring attached to block the sun. This
gave a measure of the diffuse radiation (the
portion of the global radiation that has been
scattered by the sky and clouds. An observation
window of one hour and ten minutes was used which
allowed the average and spread of the
measurements in that window to be calculated.
The window was centred on the human observation
time of 0900 UTC to allow comparison of
measurements. The fraction of diffuse to global
radiation diffuse fraction was found to be a
useful indicator for the cloud amount whereas the
ratio of global to top of atmosphere radiation
transmittance was found to give a good
indication of cloud type. It was found that
cloud height could not be distinguished using
this method, however it is thought that
downwelling longwave radiation will give enable
the cloud base temperature and therefore the
height to be deduced.
5. The Diffuse Fraction (DF)

• 1. Aims and Motivation
• To make cloud observations objective
• To deduce cloud cover using surface solar
  radiation data
• Large data set collected since 1947 (Stagg 1950)
• Large uncertainty in future climate predictions
  due to unknown consequences of cloud feedbacks
• The investigation of long term trends in cloud
  cover and there causes may reduce uncertainty in
  climate predictions.

• This graph shows
• The diffuse fraction measured by a pyranometer
  over a period of one hour and ten minutes is
  related to the cloud amount classified by a human
  observer looking at the entire sky dome.
• There are three separate regimes to the behaviour
  of the diffuse fraction. There is an aerosol
  limit ranging from 0 to 2 oktas and an overcast
  limit between 7 and 8 oktas.
• The range in which the diffuse fraction is most
  sensitive ranges from 3 to 6 oktas were there
  appears to be a linear association.

• 3. Methodology
• Cloud amount and cover is recorded by a human
  observer every morning at 0900 UTC. A photograph
  is also taken of the sky at this time to allow
  for comparison
• The diffuse and global radiation is collected as
  five minute measurements. These measurements are
  then averaged over a one hour and ten minute
  observation window, centred on the human
  observation time
• The diffuse fraction and transmittance are
  calculated over this window and compared to the
  human observations

(a) (b)
Cloud Amount

• 7. Conclusions and future work
• Surface solar radiation can be used to infer
  cloud amount when split into its components
  (direct and diffuse radiation).
• Cloud type is harder to distinguish using this
  method and only cumuliform and stratiform cloud
  were distinct enough to separate.
• This method give an hourly averaged cloud cover
  measurement which is objective.
• The cloud amount and type can be deduce at
  anytime during daylight using this method.
• Cloud height is not able to be judged using solar
  radiation. It is though that downwelling
  longwave radiation will give an indication of
  cloud base temperature and therefore height.
  This is future work.

5
4
5
6
4
6
3
7
3
1
7
1
8
9
9

• 6. Results
• Diffuse fraction is better to deduce the cloud
  amount
• Transmittance can tell us about the cloud form
  (i.e. stratus or cumulus)
• Cant use just this data to deduce cloud height

0
2
8
2
0
Figure 7 Mean plotted against the standard
deviation of (a) DF and (b) (1 T) between 0825
and 0935 with numbers indicating the cloud amount
in oktas as recorded by a human observer at 0900
UTC. Each point represents the median of all
measurements in each cloud amount category with
error bars represent one standard deviation of
the medians.
Cloud Type
(a) (b)
Cu
Cu
7 oktas stratocumulus 0900 UTC 26/10/05
References Hamblyn, R., 2001, The Invention of
Clouds. Picador, United Kingdom Monteith, J.L.
and Unsworth, M. H., 1990, Radiation
Environment. Pp. 36-57 in Principles of
Environmental Physics. Butterworth-Heinmann,
Oxford Stagg, J., 1950, Solar Radiation at Kew
Observatory Geophys Mem No 86 (first number, vol.
11). UK Met Office, HMSO Steven, M. D., and
Unsworth, M. H., 1980, Shade-ring corrections for
pyranometer measurements of diffuse solar
radiation from cloudless skies. Quart. J. R. Met.
Soc., 106, 865-872 World Meteorological
Organization, 1983, Measurement of Radiation,
Pp. 9.1-9.55 in Guide to Meteorological
Instruments and Methods of Observation. WMO,
Geneva
3 oktas cumulus 0900 UTC 20/08/05
Sc
Ci, Cc, Cs
Ci, Cc, Cs
Ac, As
Sc
StCu
St
St
Ac, As
None
St
Ac, As
None
Figure 4 Mean plotted against standard deviation
of (a) DF and (b) (1 T) between 0825 and 0935
UTC. Each point represents the median of all
measurements in each cloud type category as
classified by a human observer at 0900 UTC, with
the bars representing one standard deviation of
the median.
8 oktas stratus 0900 UTC 02/08/05
0 oktas 0900 UTC 17/08/05