A1257278317pdqVD

1
Statistical description of UV climate and
climatological maps Second EDUCE Meeting,
Bordeaux 2001 H. Slaper and P.N. den
Outer
2

• UV climatology
• Retrieve and analyse large amounts of data,
  spectral and pyranometer data. Sets
  requirements on format (data format, cosine
  correction,.)
• Construction of integrated properties daily,
  monthly, annual dosesV-dat How to deal
  with data gaps?
• Models required since UV-data exist for limited
  periods only Validation of models
  (ground-based and satellite-based)
• QA/QC-tools on irradiance scale errors

3

• DATA SUPPLEMENTATION
• Data gaps supplemented with
• RB-measurements
• fractional
• Test
• 35 days interchanged
• 1. randomly
• 2. in a row
• Data used
• 1998, skin cancer weighted
• RIVM, Bilthoven, The Netherlands
• RB cal. ?RB ? spectroradiometer

4
Months Random
Average UVsupp/UVmeas s.d. UVsupp/UVmeas s.d.
RB-supp. 1.000 0.006 1.000 0.002 Fractional
1.000 0.005 1.000 0.003
5
Extra terrestrial UV
Ozone SZA Standard atmosphere
UV-transfer model
Cloudless sky UV
1. Global Solar Irradiation, pyranometers,
WRDC 2. TOMS Reflection 3. ISCCP (?, tCLD)
UV-reduction algorithm
Cloudy sky UV
6
UV v.s. Global Solar Irradiation
7
Yearly doses
Monthly doses
Trends 51/decade, cloudless 62/decade,
cloudy sky Standard deviations in yearly
sums 1.00 0.01, cloudless 0.98 0.03, cloudy
sky
8
CORRELEATION WITH GROUND-BASED DATA
ISCCP, and TOMS versus GDF Belgium Southern
Netherlands
correlation
9
(No Transcript)
10
Cloudy sky
Cloudless sky
Percent change
Percent change
11
Cloudless sky Cloudy sky
12
(No Transcript)
13
UV v.s. Global Solar Irradiation
14
UVB 300-310 nm
UVB 300-310 nm, summer data
Clear sky model Direct related SZA-dependent
Dialy sums
Individual measurements
Global Irradiation
UVA 340-350 nm
15
Results QA/QC-absolute irradiance Gaussian
fit 95 conf. sd sd All
UVA 0.086 0.103 summer UVA 0.079 0.103 All
UVB 0.12 0.16 summer UVB 0.103 0.13 Check
threshold UVA 20 UVB 35 Thresholds
apply to all weather conditions
16
(No Transcript)
17
Conclusions 1) data gaps can be eliminated
without introducing much uncertainty 10
loss of data adds 0.5 uncertainty 2) Use of
pyranometer and ozone data is good approach to
model UV. The approach yields good correlation
with satellite data based models
next step validation at other locations using
DB 3) Agreement of model with measurements
allows for irradiance checks on the 20
level