The project has so far concentrated on two of the five objectives listed in the project description part:
1. Compilation of a ten-year (1991-2000) cloud climatology
2. Development of a new generation cloud analysis model
Concerning the objectives oriented towards BALTEX and climate model validation tasks, a new SNSB project (entitled: "Validation of cloud parameters in regional climate model simulations using NOAA AVHRR data") has been formed to deal with them more specifically.
Compilation of a ten-year (1991-2000) cloud climatology
The compilation of the cloud climatology dataset is currently in its final completion phase and the entire dataset will be available in the beginning of 2001. It is planned to store results on CD-ROMs in the HDF (HDF5) format.
Results from selected early periods of the climatology dataset has been extracted and used separately. For example, a limited comparison with the ISCCP dataset has been made (Karlsson, 1997_1) as well as an experimental model validation study where cloud forecasts of the HIRLAM model were evaluated (Karlsson, 1996_2).
For some selected months, longer period climatologies have been extracted and published (e.g. Karlsson, 1997_3 and Karlsson, 2000). An example of a six-year climotology for some selected months is shown below.
April July December
Scandinavian six-year mean (1991-1996) of cloud cover for April, July and December at 4
km resolution derived from AVHRR data (the SCANDIA model operated at SMHI).
Development of a new generation cloud analysis model
SMHI is developing a number of cloud products based on AVHRR/AMSU data from current NOAA (15 and onwards) and future Metop satellites.
As of March 2000, prototypes for the Cloud Mask, Cloud Type, and Precipitating Clouds products are ready. The prototype for the Cloud Top Temperature and Height product is under development. Examples of all four products are shown below.
Cloud Mask
Cloud Type 
CTTH
Precipitating Clouds 
Examples of the Cloud Mask, Cloud Type, Cloud Tope Temperature and Height
(CTTH) and the Precipitating Clouds products from NOAA 15 26th of September 1999 17:58
UTC. A detailed description of the colour legends used here can be found at http://www.smhi.se/saf.
Summarised description of products
A more comprehensive summary of all products, than the one given below, can be found at http://www.smhi.se/saf and a detailed description can be found in a newly issued report available at SMHI.
Cloud Mask
The Cloud Mask product attempts to delineate all absolutely cloud-free pixels in a
satellite scene with a high confidence. In addition, it will identify cloud free snow or
ice contaminated pixels when illumination allows.
The Cloud Mask scheme is a thresholding algorithm, utilising all 5/6 spectral channels of
the AVHRR/2 or AVHRR/3 sensor, NWP short range forecast data, and 1 km GIS (digital
elevation model and land use) data. The scheme makes use of off-line radiative transfer
simulations (6S and RTTOV) of cloud free atmospheres, to estimate, prior to the reception
of satellite data, the optimal thresholds valid for the given satellite scene.
The AVHRR Cloud Mask output consists of 6 category main output, a quality flag, and a
dust/volcanic plume flag. The main output is given by the 6 categories: Unprocessed (0);
Cloud free land or cloud free sea (1); Cloud contaminated (2); Cloud filled (3); Snow/Ice
contaminated (4); Unclassified.
Cloud Type
The main objective of the Cloud Type product is to distinguish between thin and opaque
clouds and provide a rough estimate of the cloud top height, and try to distinguish
between water particle clouds and ice particle clouds. In addition to this the aim is to
be able to delineate stratiform and cumuliform clouds. The highest priority is given to
the reliable identification of the major cloud categories: low, medium, high, and
semi-transparent cirrus.
The Cloud Type algorithm takes as input the Cloud Mask output, and utilise all 5/6
spectral channels of the AVHRR/2 or AVHRR/3 sensor, NWP short range forecast data, and 1
km GIS (digital elevation model and landuse) data. The algorithm distinguish different
cloud types using thresholds defined by off-line radiative transfer calculations and a
database of interactively collected training targets.
The AVHRR Cloud Type output consists of a 21 category main output, a quality flag, and a
cloud particle phase flag. For a full listing of all cloud classes see at http://www.smhi.se/saf.
CTTH
The Cloud Top Temperature and Height product will take as input the Cloud Type, and
utilise the 11 and 12 micron channels of the AVHRR sensor, NWP short range forecast data,
and 1 km GIS (digital elevation model and land use) data. To compensate for the
atmospheric attenuation above the opaque clouds, and for the semi-transparency correction,
radiative transfer calculations (RTTOVS) using NWP short range forecast output of the
vertical temperature and humidity profiles will be used.
The output will provide information on the temperature and height (in meters or hPa) of
opaque, and depending on the validation results, possibly also semi-transparent and
sub-pixel cloud tops. In addition to the main output there will be, as for the Cloud Mask
and Cloud Type, a quality flag (or processing flag). This flag tries to describe the
status of the Cloud Mask/Type input data, the status of the NWP input data, the status of
the satellite input data, the method used for the retrieval, and for the quality of the
retrieval process.
Precipitating Clouds
The objective of the Precipitating Clouds product is to support detailed precipitation
analysis for Nowcasting purposes. The focus is on the delineation of non-precipitating and
precipitating clouds for light and heavy precipitation, rather than quantifying the
precipitation rate. Particular attention will be given to the identification of areas of
light frontal precipitation.
The product will provide probability results, i.e. probabilities of precipitation
intensities in pre-defined intensity intervals. From the probabilities a categorical
estimate of precipitation intensity may be derived. It is not part of the PC product to
provide information on the type of precipitation.
The algorithm(s) to retrieve information on the presence (including rough intensity
estimations) of precipitation will be based on the Cloud Type output. The method makes use
of the spectral information of the VIS/IR of the AVHRR and the microwave information of
the AMSU-A/B. The VIS/IR the day-algorithm rely on the microphysical information available
in the 1.6 ?m and 3.7 ?m channels, and the night-time algorithm rely on the 11 and 12 ?m
channels. The AMSU-algorithm will utilise the emission and scattering signal over ocean,
and the scattering signal over land, using the AMSU-A/B 23.8, 89.0 and 150.0 GHz channels.
The two algorithms will be merged to utilise the precipitation signatures in both the
AVHRR and the AMSU data.
The PC product shall consist of a numerical value for the likelihood of precipitation
within the following four precipitation intensity classes: no precipitation (R < 0.1
mm/hr; risk for (light) precipitation: 0.1 mm/hr < R < 0.5 mm/hr; light/moderate
precipitation: 0.5 mm/hr < R < 5.0 mm/hr; intensive (convective) precipitation: 5.0
mm/hr < R.
Time schedule
According to the SAFNWC project plan an integrated software package including the four
Cloud Products based on AVHRR and AMSU data, presented here, shall be ready by January
2002. Following some basic system requirements it shall be possible to install and run
this software in an operational mode at any given National Weather Service among the
EUMETSAT member states.