Product User Manual for SAFNWC/MSG “Precipitating

Code:SAF/NWC/CDOP/SMHI/SCI/PUM/4
Product
User
Manual
for
Issue: 1.5.1 Date: 25 May 2010
SAFNWC/MSG “Precipitating Cloud”
File: SAF-NWC-CDOP-SMHI-SCI-PUM(PC-PGE04 v1.5.1)
4_v1_5_1.DOC
Page: 1/18
Product User Manual for
SAFNWC/MSG “Precipitating Cloud”
(PC-PGE04 v1.5.1)
SAF/NWC/CDOP/SMHI/SCI/PUM/4, Issue 1,
Document Revision 5.1
25 May 2010
Applicable to SAFNWC/MSG version 2010
Applicable to the following PGE:s:
PGE
Acronym Product ID
PGE04
PC
Prepared by SMHI
Product name
SAFNWC/MSG/PGE04 Precipitating Clouds
Version
number
1.5.1
Code:SAF/NWC/CDOP/SMHI/SCI/PUM/4
Product
User
Manual
for
Issue: 1.5.1 Date: 25 May 2010
SAFNWC/MSG “Precipitating Cloud”
File: SAF-NWC-CDOP-SMHI-SCI-PUM(PC-PGE04 v1.5.1)
4_v1_5_1.DOC
Page: 2/18
REPORT SIGNATURE TABLE
Function
Prepared by
Name
SMHI, Anke Thoss
Reviewed by
Pilar Fernandez
Authorised by
AEMET, NWCSAF
Manager
Signature
Date
25 May 2010
Code:SAF/NWC/CDOP/SMHI/SCI/PUM/4
Product
User
Manual
for
Issue: 1.5.1 Date: 25 May 2010
SAFNWC/MSG “Precipitating Cloud”
File: SAF-NWC-CDOP-SMHI-SCI-PUM(PC-PGE04 v1.5.1)
4_v1_5_1.DOC
Page: 3/18
DOCUMENT CHANGE RECORD
Version
Pages
CHANGE(S)
23 January 2009
2 March 2009
16
17
1.5.1d
19 April 2010
22
1.5.1
25 May 2010
18
New Document for v2009
Changes after DRI-2009:
-Corrected erroneous references and add short reference
names
-Acronym list updated
-Erroneous reference to PPS in chapter 1.3 deleted
-Applicable documents: dates and codes updated
-clarified reference to last scientific update in section 1.7
. – no scientific updates. Adapted date, issue and
revision to v. 2010
- Included reference to VS report nov. 2009
Added the full NWCSAF logotype on first page.
1.5d
1.5
Date
Code:SAF/NWC/CDOP/SMHI/SCI/PUM/4
Product
User
Manual
for
Issue: 1.5.1 Date: 25 May 2010
SAFNWC/MSG “Precipitating Cloud”
File: SAF-NWC-CDOP-SMHI-SCI-PUM(PC-PGE04 v1.5.1)
4_v1_5_1.DOC
Page: 4/18
Table of contents
1.
INTRODUCTION.........................................................................................................................................6
1.1
1.2
1.3
1.4
1.5
1.6
1.6.1
1.6.2
1.7
2.
SCOPE OF THE DOCUMENT ......................................................................................................................6
SCOPE OF OTHER DOCUMENTS ................................................................................................................6
WHO SHOULD READ THIS MANUAL .........................................................................................................6
SOFTWARE VERSION IDENTIFICATION ....................................................................................................6
DEFINITIONS, ACRONYMS AND ABBREVIATIONS ....................................................................................6
REFERENCES ...........................................................................................................................................7
Applicable Documents.......................................................................................................................7
Reference Documents ........................................................................................................................7
SCIENTIFIC UPDATES SINCE MSG VERSION 2009...................................................................................8
DESCRIPTION OF THE PRODUCT ........................................................................................................9
2.1
GOAL OF THE PC PRODUCT .....................................................................................................................9
2.2
OUTLINE OF THE PRECIPITATING CLOUD ALGORITHM .............................................................................9
2.2.1 General algorithm design..................................................................................................................9
2.2.2 Data sources for Precipitationg Clouds............................................................................................9
2.2.3 Graphical overview of the Precipitating Clouds product (PGE04) ................................................10
2.2.4 Description of the output.................................................................................................................11
2.2.5 Statistics file ....................................................................................................................................12
3.
IMPLEMENTATION OF THE PRODUCT ............................................................................................13
4.
INPUTS AND CONFIGURABLE PARAMETERS ................................................................................13
4.1
4.2
5.
LIST OF INPUTS .....................................................................................................................................13
CONFIGURABLE PARAMETERS ..............................................................................................................13
VALIDATION.............................................................................................................................................13
5.1
SUMMARY OF VALIDATION RESULTS ....................................................................................................13
6.
KNOWN PROBLEM AREAS AND LIMITATIONS .............................................................................13
7.
EXAMPLE OF PRODUCT VISUALISATION.......................................................................................13
Code:SAF/NWC/CDOP/SMHI/SCI/PUM/4
Product
User
Manual
for
Issue: 1.5.1 Date: 25 May 2010
SAFNWC/MSG “Precipitating Cloud”
File: SAF-NWC-CDOP-SMHI-SCI-PUM(PC-PGE04 v1.5.1)
4_v1_5_1.DOC
Page: 5/18
List of Tables and Figures
Table 1: List of Applicable Documents.............................................................................................7
Table 2: List of Referenced Documents ............................................................................................7
Table 3: Quality flags of PGE04 .....................................................................................................12
Figure 1: schematic overview over the Precipitating Clouds product
10
Figure 2: 200901241200 precipitating clouds product over MSG-N, configured for day algorithm.
Dark green hues present precipitation likelihood classes 10%-20%,light green 30%, yellow40%
and orange/red hues 50% and higher.
12
Figure 3: Likelihood of rain from PC product versus observed rain frequency. Cloud type
dependent tuning on French gauge data.Left: independent validation against Hungarian gauge data
2004, right: performance on dependent French gauge data set 2004. verified against 30 min
averages in gauge data. Blue:high and very high clouds, red: medium level clouds,green: thick
cirrus,cyan: cirrus over lower clouds.
13
Figure 4: Likelihood of rain versus observed rain frequency. Same as Figure 3 but for night time.
13
Figure 5: Example of the precipitating clouds product with a change from day to night algorithm
diagonally over the British isles (northward night algorithm, south day algorithm).Please note
typical features: precipitation area more spread out for night time algorithm, less detailed features
and no high precipitation likelihood. At high satellite viewing angles the product becomes
unreliable as seen by high precipitation likelihood at the rim of the Meteosat disk.(do not use for
satellite viewing angles exceeding 60 degrees).
13
Code:SAF/NWC/CDOP/SMHI/SCI/PUM/4
Product
User
Manual
for
Issue: 1.5.1 Date: 25 May 2010
SAFNWC/MSG “Precipitating Cloud”
File: SAF-NWC-CDOP-SMHI-SCI-PUM(PC-PGE04 v1.5.1)
4_v1_5_1.DOC
Page: 6/18
1. INTRODUCTION
The Eumetsat “Satellite Application Facilities” (SAF) are dedicated centres of excellence for
processing satellite data, and form an integral part of the distributed EUMETSAT Application
Ground Segment ( http://www.eumetsat.int ). This documentation is provided by the SAF on
Support to Nowcasting and Very Short Range Forecasting, SAFNWC. The main objective of
SAFNWC is to provide, further develop and maintain software packages to be used for
Nowcasting applications of operational meteorological satellite data by National Meteorological
Services. More information can be found at the SAFNWC webpage, http://www.nwcsaf.org . This
document is applicable to the SAFNWC processing package for Meteosat satellites
meteorological satellites, SAFNWC/MSG.
1.1 SCOPE OF THE DOCUMENT
This document is the Product User Manual for the SAFNWC MSG Precipitating Clouds product. The
document describes how to use the product after installation. It is meant to support the interpretation as
well as describe the possibilities and limitations.
1.2 SCOPE OF OTHER DOCUMENTS
The algorithm used in the Precipitating Clouds Product is described in more detail in the
corresponding Algorithm Theoretical Basis document [AD. 6].
Validation of the algorithm is detailed in the Validation report for Precipitating Clouds [AD. 1].
Instructions how to install, configure and execute the software are given in the Software User
Manual for NWCSAF MSG Package [AD. 2].
The Interface Control Documents [AD. 3] (for the External and Internal Interfaces of the
SAFNWC/MSG) and [AD. 4](MSG Output Product Format Definition) detail the input and output
data format for the SAFNWC/MSG software.
1.3 WHO SHOULD READ THIS MANUAL
This document is intended for the end-user, i.e. the forecaster.
For the person in charge of building and installing the MSG software package, thus the sys-admin
we refer to the Software User Manual ([AD. 2]), which will of course also be relevant for the
science-admin.
For the person interested in the algorithms in detail we refer to the Algorithm theoretical Basis
Document[AD. 6].
1.4 SOFTWARE VERSION IDENTIFICATION
This document describes the algorithms implemented in the PGE04 version v1.6 of the 2010
SAFNWC/MSG software package delivery.
1.5 DEFINITIONS, ACRONYMS AND ABBREVIATIONS
Acronym
CDOP
Explanation
Continuous Development and
Operational Phase
Acronym
CM
CT
Explanation
Cloud Mask (also PGE01)
Cloud Type (also PGE02)
Code:SAF/NWC/CDOP/SMHI/SCI/PUM/4
Product
User
Manual
for
Issue: 1.5.1 Date: 25 May 2010
SAFNWC/MSG “Precipitating Cloud”
File: SAF-NWC-CDOP-SMHI-SCI-PUM(PC-PGE04 v1.5.1)
4_v1_5_1.DOC
Page: 7/18
Acronym
CTTH
EUMETSAT
FAR
FOV
HDF5
IR
LUT
MSG
NIR
NORDRAD
NWP
PC
Explanation
Cloud Top Temperature, Height
and Pressure (also PGE03)
European Organisation for the
Exploitation of Meteorological
Satellites
False Alarm Rate
Field Of View
Hierarchical
Data
format
version 5
Infrared
Look-Up-Table
Meteosat second generation
Near Infrared
Nordic Weather Radar Network
Numerical Weather Prediction
Precipitating
Cloud
(also
PGE04)
Acronym
PCPN
PGE
PI
POD
POFD
RGB
SAF
SAFNWC
SEVIRI
SMHI
SW
TOA
USGS
VIS
Explanation
Precipitation
Process Generating Element
Precipitation Index
Probability Of Detection
Probability Of False Detection
Red Green Blue
Satellite Application Facility
Satellite Application Facility
for support to NoWcasting
Imager onboard MSG satellites
Swedish Meteorological and
Hydrological Institute
SoftWare
Top Of Atmosphere
U.S. Geological Survey
Visible
For a list of SAF-acronyms see also [RD.1].
1.6 REFERENCES
1.6.1 Applicable Documents
Reference
[AD. 1]
[AD. 2]
[AD. 3]
[AD. 4]
[AD. 5]
[AD. 6]
[AD. 7]
Title
Validation Report for Precipitating Clouds (PCPGE04v1.4)
Software User Manual for the SAFNWC/MSG
Application: Software Part
Interface Control Document for the External
and Internal Interfaces of the SAFNWC/MSG
SAFNWC/MSG Output Product Format
Definition
Architectural Design Document for the
SAFNWC
Algorithm Theoretical Basis Document for
"Precipitating Clouds" (PC-PGE04 v1.5.1)
Cross-Verification of the Rapid
Development Thunderstorm and the
Precipitation Products of the
Nowcastion and Vert Short Range
Forecasting SAF
Code
SAF/NWC/CDOP/SMHI/SCI/VR/01
Vers
1.4
Date
19/11/07
SAF/NWC/CDOP/INM/SW/SUM/2
4.0
16/02/2009
SAF/NWC/CDOP/INM/SW/ICD/1
4.0
2010
SAF/NWC/CDOP/INM/SW/ICD/3
4.0
2010
SAF/NWC/CDOP/INM/SW/AD/1
4.0
2010
1.5.1
25/05/2010
SAF/NWC/CDOP/SMHI/SCI/ATBD/
04
15/11/2009
Visiting scientist report by
Eszter Lábó, Mária Putsay,
Zsófia Kocsis and Ildikó
Szenyán
Table 1: List of Applicable Documents
1.6.2 Reference Documents
Reference
[RD.1]
[RD.2]
Title
Code
Vers
Date
The Nowcasting SAF Glossary
SAF/NWC/CDOP/INM/MGT/GLO
1.3
10/11/2009
SAFNWC Product Requirements Document
SAF/NWC/INM/MGT/PRD
1.0
28/07/2009
Table 2: List of Referenced Documents
Product User Manual for CodeSAF/NWC/CDOP/SMHI/SCI/PUM/4
SAFNWC/MSG “Precipitating Issue: 1.5.1 Date: 25 May 2010
File:SAF-NWC-CDOP-SMHI-SCI-PUMCloud” (PC-PGE04 v1.5.1)
4_v1_5_1.DOC
Page:
1.7 SCIENTIFIC UPDATES SINCE MSG VERSION 2009
No scientific updates have been implemented since product version 1.3, released spring 2007.
8/18
Product User Manual for CodeSAF/NWC/CDOP/SMHI/SCI/PUM/4
SAFNWC/MSG “Precipitating Issue: 1.5.1 Date: 25 May 2010
File:SAF-NWC-CDOP-SMHI-SCI-PUMCloud” (PC-PGE04 v1.5.1)
4_v1_5_1.DOC
Page:
9/18
2. DESCRIPTION OF THE PRODUCT
2.1 GOAL OF THE PC PRODUCT
Goal of the PC product is to give a first guideline to the forecaster where to expect precipitation,
especially for areas where no radar data is available. The product provides the probability of
precipitation for each Meteosat pixel.
Since the coupling of radiances from visible and infrared channels with precipitation is rather
weak, large areas are marked as potentially precipitating (more than 10% precipitation likelihood).
The skill the derive stratiform precipitation is limited and potential precipitation area is
overestimated, but indicating low likelihood. Strong convective precipitation can be better
estimated from METEOSAT data than rain from stratiform precipitation, and the NWCSAF
convective rain rate product and the rapidly developing thunderstorm product can be consulted for
more detailed analysis of severe convection.
2.2 OUTLINE OF THE PRECIPITATING CLOUD ALGORITHM
2.2.1 General algorithm design
The precipitating clouds product gives the total likelihood of precipitation without attempting to
estimate intensity. To derive the likelihood of precipitation, a precipitation Index PI is constructed
from those IR and visible spectral features which are most correlated with precipitation. The
precipitation likelihood for each value of the PI is determined statistically by comparison with
collocated precipitation measurements. For the tuning of the current algorithm version French
gauge network measurements for one year of data were used.
In the calculation of the PI special attention has been given to spectral features in the visible,
which implicitly contain information on cloud microphysical properties at the cloud top, such as
effective radius and cloud phase. The algorithm employed is cloud type dependent in the sense
that mapping from PI to precipitation likelihood makes use of cloud type dependent lookup tables.
For the PI calculation a day and a night version exists, where the night version only makes use of
IR channels not influenced by sunlight.
2.2.2 Data sources for Precipitationg Clouds
•
Meteosat visible and IR channels:
Daytime: vis0.6, NIR1.6, IR3.9, IR6.2, IR7.3, IR10.8, IR12.0
Nighttime: IR6.2, IR7.3, IR10.8, IR12.0
•
Cloud type product
•
NWP surface temperature
Product User Manual for CodeSAF/NWC/CDOP/SMHI/SCI/PUM/4
SAFNWC/MSG “Precipitating Issue: 1.5.1 Date: 25 May 2010
File:SAF-NWC-CDOP-SMHI-SCI-PUMCloud” (PC-PGE04 v1.5.1)
4_v1_5_1.DOC
Page:
2.2.3 Graphical overview of the Precipitating Clouds product (PGE04)
Figure 1: schematic overview over the Precipitating Clouds product
10/18
Product User Manual for CodeSAF/NWC/CDOP/SMHI/SCI/PUM/4
SAFNWC/MSG “Precipitating Issue: 1.5.1 Date: 25 May 2010
File:SAF-NWC-CDOP-SMHI-SCI-PUMCloud” (PC-PGE04 v1.5.1)
4_v1_5_1.DOC
Page:
11/18
2.2.4 Description of the output
The precipitating clouds product gives the likelihood of precipitation:
•
Class 1: total precipitation likelihood for rain > 0.1 mm/h
•
Class2: obsolete, set to 0
The likelihood is given in intervals of 10%:
0% to
5% →
0%
> 5% to 15% → 10%
> 15% to 25% → 20%
> 25% to 35% → 30%
> 35% to 45% → 40%
> 45% to 55% → 50%
> 55% to 65% → 60%
> 65% to 75% → 70%
> 75% to 85% → 80%
> 85% to 95% → 90%
> 95% to 100% → 100%
Please pay attention to that the value in the HDF5 dataset has to be multiplied by the scaling
factor 10, as specified in the HDF file to arrive at the correct likelihood. The product is supplied in
HDF5 format and has the same resolution as in the original SEVIRI image.
Special count = 15= 150% used when no data value is available. The forecaster is likely to receive
the product displayed as an image, similar to Figure 2.
Using the 20% class as a threshold for precipitation detection gives a fairly save estimate of
possible precipitation, although some light to moderate precipitation migh be missed (deep green
in Figure 2). Using 30% as theshold (light green in Figure 2) provides usually a subjectively better
fit to radar data, but more real precipitation remains undetected while there is still a slight
overestimation of precipitation area.
2.2.4.1 Flags
The product contains also quality information in a separate field. The quality information is
indicating under which circumstances the PC product was derived. Except for the field indicating
whether solar channels were used, it should not be of much interest to the forecaster, and most
likely it will not be required to visualize this information.
Product User Manual for CodeSAF/NWC/CDOP/SMHI/SCI/PUM/4
SAFNWC/MSG “Precipitating Issue: 1.5.1 Date: 25 May 2010
File:SAF-NWC-CDOP-SMHI-SCI-PUMCloud” (PC-PGE04 v1.5.1)
4_v1_5_1.DOC
Page:
Bit #
12/18
Meaning of he bit – 1/0
0
Processed/non-processed
1
MSG channels missing/not missing
2
CT used/not used
3
MSG solar channels used/not used
4
MSG land/no land
5
High terrain/no high terrain
6
NWP data missing/not missing
7
MSG cloud mask low quality/no low quality
Table 3: Quality flags of PGE04
The quality information is indicating under which circumstances the PC product was derived.
2.2.5 Statistics file
The statistics file is an ASCII file summarising the distribution of probabilities over the complete
region. It can be easily used for verification whether 2 runs are identical. Files may also be used to
easily accumulate statistics on general algorithm performance.
Figure 2: 200901241200 precipitating clouds product over MSG-N, configured for day algorithm.
Dark green hues present precipitation likelihood classes 10%-20%,light green 30%, yellow40%
and orange/red hues 50% and higher.
Product User Manual for CodeSAF/NWC/CDOP/SMHI/SCI/PUM/4
SAFNWC/MSG “Precipitating Issue: 1.5.1 Date: 25 May 2010
File:SAF-NWC-CDOP-SMHI-SCI-PUMCloud” (PC-PGE04 v1.5.1)
4_v1_5_1.DOC
Page:
13/18
3. IMPLEMENTATION OF THE PRODUCT
The implementation is described in Software User Manual ([AD. 2]). Interesting for the forecaster
is that the region is configurable. However auxiliary data for a new region has to be compiled
beforehand. The product generation is usually scheduled automatically by the task manager.
Product User Manual for CodeSAF/NWC/CDOP/SMHI/SCI/PUM/4
SAFNWC/MSG “Precipitating Issue: 1.5.1 Date: 25 May 2010
File:SAF-NWC-CDOP-SMHI-SCI-PUMCloud” (PC-PGE04 v1.5.1)
4_v1_5_1.DOC
Page:
14/18
4. INPUTS AND CONFIGURABLE PARAMETERS
4.1 LIST OF INPUTS
Please note that the precipitating cloud product (PGE04) requires the cloudtype product as input
(PGE02), and for that even the Cloudmask product has to be run (PGE01).
•
•
•
•
NWP surface temperature
Sun zenith, satellite view zenith, and sun-satellite view relative azimuth difference angle
Output from Cloud Type
Seviri data
4.2 CONFIGURABLE PARAMETERS
The Precipitating Cloud product has been designed to allow a full configuration and flexibility to
update/tune the algorithm without having to modify the code. However most of these configurable
parameters are only of interest to the developer. The only configurable parameters potentially of
interest to the users are the configuration of when to switch from day- to nighttime scheme, and
that it is in principle possible to define which cloud types are treated as potentially raining.
•
The default configuration is that the night time algorithm is activated when the sun zenith
angle is greater than 80 degrees. It is possible to configure the product to only use the
night algorithm by setting the sun zenith angle threshold to 0 in the algorithm
configuration file. This would avoid discontinuities in the product at the day/night
deliminator on the cost of degrading performance during day time.
•
In principle it is possible to configure which cloud classes are treated as potentially
raining. Please consult the NWCSAF helpdesk before changing the validated default
configuration.
The possible configurable parameters are described in the Software User Manual [AD. 2].
Product User Manual for CodeSAF/NWC/CDOP/SMHI/SCI/PUM/4
SAFNWC/MSG “Precipitating Issue: 1.5.1 Date: 25 May 2010
File:SAF-NWC-CDOP-SMHI-SCI-PUMCloud” (PC-PGE04 v1.5.1)
4_v1_5_1.DOC
Page:
15/18
5. VALIDATION
5.1 SUMMARY OF VALIDATION RESULTS
The PC product can be validated against co-located radar data, synop current weather observations
or rain gauge data. For more information on product validation see validation reports [AD. 1] and
[AD. 6]. When verifying likelihood results of the PC product, it is important to somehow quantify
the algorithm performance and give guidance to answer the question whether it is raining or not. It
is important to understand that a simplified “categorical estimate”, which has been derived from
the likelihood distribution, degrades the product on the one hand (no “fair” comparison) but, on
the other hand, makes it more practical to use for the forecaster.
A simple way to convert likelihood estimates into easily verifiable estimates of precipitation is to
set a threshold for rain according to algorithm performance. Which threshold of total precipitation
likelihood does best divide the precipitating from the non-precipitating events? (Usually 20% or
30% of total precipitation likelihood!)
The performance of this “hard-clustering” is verified using contingency tables. Evaluating the
performance at different threshold levels gives also an overview of how closely assigned
probability values match real occurrence of rain, as illustrated in Figure 3 and Figure 4.
French dataset
Cloud type dependent tuning
Cloud type dependent tuning
Day algorithm
Day algorithm
70
80
70
60
50
40
30
20
10
0
Observed rain freq.[%]
Observed rain freq.[%]
Hungarian dataset
60
50
40
30
20
10
0
0% 10% 20% 30% 40% 50% 60% 70%
0% 10% 20% 30% 40% 50% 60% 70%
Likelihood of rain [%]
Likelihood of rain [%]
CT=11-14
CT=17
CT=9-10
CT=18
CT=11-14
CT=17
CT=9-10
CT=18
Figure 3: Likelihood of rain from PC product versus observed rain frequency. Cloud type
dependent tuning on French gauge data.Left: independent validation against Hungarian gauge
data 2004, right: performance on dependent French gauge data set 2004. verified against 30 min
averages in gauge data. Blue:high and very high clouds, red: medium level clouds,green: thick
cirrus,cyan: cirrus over lower clouds.
Product User Manual for CodeSAF/NWC/CDOP/SMHI/SCI/PUM/4
SAFNWC/MSG “Precipitating Issue: 1.5.1 Date: 25 May 2010
File:SAF-NWC-CDOP-SMHI-SCI-PUMCloud” (PC-PGE04 v1.5.1)
4_v1_5_1.DOC
Page:
16/18
Hungarian dataset
French dataset
Night algorithm dispersed on cloud type
Night algorithm dispersed on cloud type
70
Observed rain freq.[%]
Observed rain freq.[%]
70
60
50
40
30
20
10
60
50
40
30
20
10
0
0
0% 10% 20% 30% 40% 50% 60% 70%
0% 10% 20% 30% 40% 50% 60% 70%
Likelihood of rain [%]
Likelihood of rain [%]
CT=11-14
CT=17
CT=9-10
CT=18
CT=11-14
CT=17
CT=9-10
CT=18
Figure 4: Likelihood of rain versus observed rain frequency. Same as Figure 3 but for night time.
Algorithm performance can be summarized as follows:
•
At the 20% detection threshold day and night algorithms perform almost equally
well, whereas the day algorithm clearly exhibits more skill at the 30% threshold than
the night algorithm. 20%can be used as a kind of hardclustering threshold for
precipitation, but thresholding at 30% generally gives a better subjective fit to radar
precipitation areas.
•
Day and nighttime algorithms exhibit different characteristics, and discontinuities at the
day/night deliminator are apparent (see Figure 5). With just using IR channels at night,
there are less areas assigned high precipitation likelihood and the precipitation areas are
less defined at night time. For the night time algorithm precipitation occurrence is more
strongly overestimated in winter, in summer more actual precipitation is missed. Both at
20% and 30 % threshold precipitation occurrence is overestimated.
•
The work with separating cloud types has shown that:
Cloud type class 9-10 (medium level cloud) precipitation is overestimated at 20
percent detection level.
Cloud type class 17 and 18 (thick cirrus and cirrus over lower cloud) give bad
results overall.
Cloud type class 11-14 (high and very high cloud) seems to be the easiest to
handle
Considering cloud low clouds (CT 8) as possibly precipitating might be
considered in the following versions
Product User Manual for CodeSAF/NWC/CDOP/SMHI/SCI/PUM/4
SAFNWC/MSG “Precipitating Issue: 1.5.1 Date: 25 May 2010
File:SAF-NWC-CDOP-SMHI-SCI-PUMCloud” (PC-PGE04 v1.5.1)
4_v1_5_1.DOC
Page:
17/18
6. KNOWN PROBLEM AREAS AND LIMITATIONS
•
The current version of the product contains a certain dependence on sun zenith angle.
•
There is also a clear jump in algorithm performance between day and night algorithm,
which cannot be totally avoided.
•
The product degrades considerably at high viewing angles and use for viewing angles
greater than 60 degrees is not recommended.
•
The algorithm does currently not detect any precipitation from low clouds
Product User Manual for CodeSAF/NWC/CDOP/SMHI/SCI/PUM/4
SAFNWC/MSG “Precipitating Issue: 1.5.1 Date: 25 May 2010
File:SAF-NWC-CDOP-SMHI-SCI-PUMCloud” (PC-PGE04 v1.5.1)
4_v1_5_1.DOC
Page:
18/18
7. EXAMPLE OF PRODUCT VISUALISATION
Figure 5: Example of the precipitating clouds product with a change from day to night algorithm
diagonally over the British isles (northward night algorithm, south day algorithm).Please note
typical features: precipitation area more spread out for night time algorithm, less detailed
features and no high precipitation likelihood. At high satellite viewing angles the product
becomes unreliable as seen by high precipitation likelihood at the rim of the Meteosat disk.(do not
use for satellite viewing angles exceeding 60 degrees).