4.2 Review of data collected in each test site

A policy of routine data acquisition was adopted for the primary satellite derived data, such as synthetic aperture radar (SAR), along track scanning radiometer (ATSR), Advanced Very High Resolution Radiometer (AVHRR) and the Sea-viewing Wide Field of view Sensor (SeaWiFS). Other data were also important to the project, including data from other satellites and sensors, as well as conventional data from buoys, ships and other sources based around sea level measurement.

4.2.1 Baltic Sea

A number of data sets have been collected over the Baltic Sea test site. Apart from several satellite scenes, meteorological data are also available, plus in situ data from the Finnjet ferry. Depending on data type, origin and pre-processing level, the different data sets have been processed according to their intended use.

The bulk of the satellite data are comprised of NOAA AVHRR scenes received locally at Stockholm University. During the first two years of the project, approximately 370 GB have been received, archived and processed in a standardised way. Apart from the archiving of all raw data for later retrieval and processing on demand, several processing steps have been performed for the areas defined by the Clean Seas project. Four scenes per day were processed and interpreted for the occurrence of any pollutant, algal bloom, sediment outflow or other anomalies, such as distinct temperature fronts within the Baltic test site. For the North Sea and Gulf of Lion areas one scene per day per area were processed. All the AVHRR data within the project sites were converted to the agreed HDF format for easy accessibility by the other project partners. Products based on the AVHRR data include annotated interpreted images, either channel 1 (visible 0.58-0.68 µm) or channel 4 (thermal infrared 10.5-11.5 µm). Other ‘general’ products are quick-looks of all received scenes and sea surface temperature (SST) images of the Baltic Sea. During the cyanobacterial bloom season, images showing the surface accumulations in the Baltic Proper, Gulf of Finland and the Bothnian Sea are also available.

Since May 1998 regular reception and processing of SeaWiFS data has also been carried out at Stockholm University. So far all received data have been in delayed (encrypted) mode, i.e. it is not possible to process data until 2-3 weeks after reception. All data are archived in both encrypted form, immediately after the satellite pass, and decrypted, a few weeks later. Apart from this storage, data are also converted into level 1a and are available by ftp, in agreement with NASA. The SeaWiFS data are also processed and interpreted for the occurrence of pollutants etc, but only after the 2-3 weeks delay period. An exception to this was during the 1999 cyanobacterial bloom, when data were received from the Dundee receiving station in near real time. This near real time reception, processing and interpretation was performed even though local data were received (in delayed mode). The reason for this dual data reception was the intense cyanobacterial bloom in the Baltic and ongoing in situ sampling simultaneously.

The ERS-2 SAR data being collected over the Baltic Sea have primarily been used for oil spill detection, but in a case study the radar images are also used for algal bloom monitoring in conjunction with several other satellite data sources. During the first two project years 220 SAR scenes have been processed over the Baltic Sea according to the pre-defined frames. The processing and interpretation of the SAR data has been done at the Institut für Meereskunde, University of Hamburg. A more detailed description of the oil spill detection and statistics based on SAR data can be found in section 5.3.1.

For the special case study on ‘SAR assisted algal bloom monitoring’ (section 5.1) both Landsat-5 TM data and WiFS from the IRS-1C satellite have been used together with AVHRR and SAR.

For the comparison between interpretations of cyanobacterial surface accumulations from NOAA AVHRR imagery and in situ observations, data from the Finish ferry ‘Finnjet’ have been used. The Finnjet ferry plies regularly between Helsinki and Travemünde while it automatically measures salinity, temperature and chlorophyll. These data have been made available through the Finnish Institute of Marine Research in Helsinki. A description of the comparison can be found in section 4.4.

Data from some meteorological and oceanographic Swedish stations in the Baltic are also available and have been used to compare observed events with meteorological conditions. Figure 4-4 shows the positions of the selected stations. Parameters observed are (varies with station); air and sea surface temperature, humidity, wind speed and direction, precipitation, sea level, sun shine duration and irradiance.

4.2.2 North Sea

ATSR

The ATSR data were obtained from ESA through PI status on ERS-2.

Initially all scenes which touched on the regions of interest were ordered, but the work load in processing by ESA was too large, and the order was cut to only include those data which covered the central areas of interest. Although both GBT (gridded brightness temperature for each channel) and GSST (gridded sea surface temperature) data were ordered for all scenes, the GSST data were not available for the first 18 months of the project. Once the GSST data were available, no more GBT data were received. We have GSST data for almost all scenes for which we received GBT but the lastest GBT data received were for 6 July 97. Unfortunately this leaves us with no visible channel data (only available in the GBT product) for comparison with the colour sensors.

The total number of scenes ordered and received for each test region are given in Table 4-1 (Kattegat scenes are included in both the North Sea and Baltic totals). The "Total Scenes" is the number of scenes for which we have GBT, or GSST, or both GBT and GSST. Also given are the first and last dates for which data have been received, although not all scenes available between these dates have been received. Less than 80% of the ordered GSST scenes were received, and only 25-30% of the GBT scenes.

AVHRR

AVHRR data were received at Stockholm University and one pass per day was processed and made available on the web for rapid viewing. All received passes were archived and distributed on CDROM, approximately one month per CD. The AVHRR data provided the most frequent and continuous data set for the North Sea region and was chosen as the ‘base’ set for selection of test periods.

Colour

A single OCTS image is available for 1 April 1997, although the source data for this image were never released by the Japanese space agency. Only 3 MOS scenes are available, for 14, 18 and 19 August 1997. SeaWiFS data are available from September 1997 and a regular time series was produced by JRC from the Dundee satellite receiving station. A good time series of images is available from 10 May 1998-1 June 1998.

Ancillary Data

For the test periods, and some additional periods under consideration for test studies, wind data collected at several coastal and offshore sites were obtained from RIKZ (Dutch Hydrographic Institute). For the modelling studies, volumes for the Rhine Outflow at several strategic locations were also obtained from RIKZ. Sediment characteristics for the modelling were also obtained from the same source. These data are held at SOC and were passed to ACRI for modelling purposes.

Sea Level data were obtained to use as boundary conditions for the model. These data were obtained directly from RIKZ by ACRI, and are held at ACRI.

In situ SST from the offshore stations were obtained for several periods overlapping with ATSR observations to confirm the accuracy of the ATSR data in this region. These data are held at SOC.

4.2.3 Gulf of Lion

The data collected for the Gulf of Lion test site consists of three sets: the general information, the satellite data, and the in situ experiments. The general information consists of the general elements describing the situation of the studied area. Most sources are cartographic, or provided by specific studies. The purpose is to describe synthetically the general conditions for the occurrence of a given polluting event, regarding both the risk of an event and the assessment of the impact in terms of sensitivity mapping. The elements relative to the risk include the industrial activity on shore and offshore, the ship routes, and the description of the possible transport mechanisms for a surface pollutant (climatology for the surface mean currents, swells, and winds). The elements for sensitivity mapping are relative to both physical, biological and socio-economical aspects. Data from previous studies acquired by Ecole des Mines for UNEP, from Ifremer, or from local authorities were used to build maps, at five scales from the global basin scale to local the scale for the description of specific sensitive areas. Figure 4-3 shows the map of the risk at the second scale, also named the sub-basin scale. Maps have been placed on the project web-site.

The collected satellite information has been of three types: ERS-2 SAR data have been continuously processed at UHSO from December 1996 to November 1998 for an area covering the Tarragona oil exploitation basin, the Barcelona coastal effluent, the Rhône river plume, and the major ship routes. After the success of the ship route monitoring achieved in 1996-1997, it was decided to extend coverage to the Marseille-Gibraltar and Marseille-Skikda routes, including night passes. The total coverage consists of a database of 303 SAR frames, 7 for 1996, 152 for 1997, and 144 in 1998. 100% of the acquired frames have been placed in the website, due to the high quality of the data, and the ability of SAR to acquire an image regardless of the observation conditions (day/night, or clouds). The AVHRR data were acquired at Dundee, and continuously processed at Ecole des Mines, for the full years 1995, 1996, 1997, and 1998. The choice of such an extended time coverage was made to complete the gap existing between previous studies and the present one, for a climatological purpose. A total of over 2920 processed frames, 277 (9.5%) have been kept for the study (19 for 1995, 72 for 1996, 100 for 1997, 86 for 1998), which is slightly smaller than the commonly accepted rate for the area (10%), regarding the general cloud conditions. ATSR data processed at SOC have also been made available for the full period of the study. SEAWiFS data have also been processed at JRC, and progressively been made available, after the launch of the sensor in August 1997. Due to the cloud cover conditions and atmospheric transparency during the period on, only a few images have been made available.

In situ data of surface water optical properties, and biological activity have been collected during the summers of 1997, 1998, and 1999, through the the Delphis campaigns. Each time, a set of 350 measurement points could have been managed, with the acquisition of SAR, AVHRR and SEAWiFS images, for 1998 and 1999. The results and the description of the campaigns have been made available on the project web-site.