R. Santoleri (1), G.L. Liberti (1), D’Ortenzio (1,2), G.L. Volpe (1,2), S. Marullo (1, 3), R. Sciarra (1),

Istituto delle Scienze dell’Ambiente e del Clima – ISAC, (formerly Istituto di Fisica dell’Atmosfera -IFA) – CNR, Roma

Stazione Zoologica A. Dorhn – Napoli

Ente Nazionale Energia e Ambiente – ENEA, Roma

SeaWiFS historical data processing shift sul paragrafo corrispondente
SeaWiFS historical geophysical data quality control “
SeaWiFS aerosol products validation “
Remote sensing of phytoplankton pigments at the basin scale “

 

SeaWiFS historical data processing

We selected as historical year the year 2000 in order to have continuity between the historical and the experimental phase SeaWiFS data. All the available level 1A SeaWiFS data (more 800 passages) covering the Mediterranean Sea for the year 2000 have been recovered from the IFA satellite archive and processed with the last version of the SeaDas software v.4.0B (Baith et al., 2001) available from NASA, which uses the new calibration coefficients (re-processing # 3).
The processing was made with the following options which are those adopted for the standard NASA products:

Multi-scattering atmospheric corrections

Aerosol model selection based on the analysis of the 765 and 865 nm radiances

Siegel (Siegel et al., 2000) NIR iterations

OC4v4 algorithm (O’Reilly et al., 2000) for chlorophyll computation

Only land and straylight flags were activated.

We decided to produce up to Level 2 the following 27 geophysical products (link a ADIOS Data): chlorophyll concentration, normalized living radiances at 412, 443, 490, 510, 555, 670, 865 nm, diffuse attenuation coefficient at 490 nm, calibrated Top of the Atmosphere radiance at 412, 443, 490, 510, 555, 670, 765, 865 nm, aerosol optical thickness at 670 nm, Ångstrom coefficient between 670 and 870, aerosol index, aerosol model types and other related parameters. Moreover, all the 24 Level 2 quality flags containing information on clouds, case 1/2 water, high aerosol concentration, etc. were also produced. Each Level 2 data file (300 Mbyte) was produced in HDF format and stored on tape. Level 2 geophysical products were successively remapped on the geographical window of covering the area (42-49N, 10W-42.3E).
Then, daily composite and monthly maps of the principal geophysical parameters (e.g. aerosol optical thickness at 670 nm, Ångstrom coefficient between 670 and 870, aerosol model types) were produced. In figure 1 an example of the products are presented.

 

SeaWiFS historical geophysical data quality control

A first type of evaluation based on a subjective analysis of each single orbit atmospheric product map and monthly and seasonal composites was performed. From such an analysis it is expected to identify evident algorithm behaviours such as the occurrence of clearly not-physical features/values, resulting e.g. from calibration, observation geometry and cloud screening. To perform such analysis each the aerosol optical thickness at 670 nm (τ) and the Ångstrom coefficient between 670 and 875 nm (α) maps were visualized on the screen and comments were reported on paper. From the analysis of images like figure 1 and b these images a few interesting features emerge:

 

SeaWiFS aerosol products validation

The validation of the SeaWiFS aerosol products over the Mediterranean Basin against in situ land observation has been started analysing the year 2000 ADIOS products. The validation data set is based on observations from AERONET (Holben et al., 1998) ground-based CIMEL sun-photometers measurements made in sites around the Mediterranean within 100 km from the coast. The following sites were selected: Avignon, El Arenosillo, Erdemli, Lampedusa, Nes Ziona, Oristano, Venezia, Nea Michaniona (figure2c).
CIMEL-AERONET instruments and the SeaWiFS radiometer have only three common wavelengths: 440, 670 and 875 nm. Therefore, as a preliminary analysis, aerosol optical thicknesses at 670 nm (τ) and the Ångstrom coefficient between 670 and 875 nm were compared.
In order to perform a validation of instantaneous and nearly co-located measurements a match-up data base was constructed. The matching criteria have been determined by analysing spatial and temporal variability characteristics of the aerosol products.
Fig. 2 shows the comparison of match-up results obtained using ground truth data in a +/- 0.5 hr time window from the satellite overpass and 40 km distance between the satellite and ground truth data. A relatively good agreement for the aerosol optical thickness is found. On the contrary, the comparison of α shows a very poor agreement between AERONET in situ and SeaWiFS estimations. On the other hand, part of the relatively large in situ values of α come from the Oristano station, which we suspect to have a calibration problem which produces larger values of α at local noon.
The evident lack of aerosol models allowing values larger than 1.7 is due to the low representativeness of the aerosol models on which is based the SeaWiFS aerosol properties retrieval. Such a result may have a strong impact on the atmospheric correction for the retrieval of the water-leaving radiances at shorter wavelengths. It seems that the SeaWiFS candidate aerosol models used in the standard NASA atmospheric correction procedure are somehow too ‘maritime’. In fact, the Mediterranean Sea is relatively enclosed basin, with low wind intensity regimes and under the influence of strong continental-like, antropogenic and natural, sources of aerosols. This hypothesis is confirmed by the analysis of the aerosol model types selected for the atmospheric correction which results in the selection of tropospheric aerosols for polluted area. Providing a more adequate local aerosol model, that could be defined on the basis of recent literature and observations, SeaWiFS, with its high radiometric sensitivity in the low reflectance range and relatively good ground resolution, appears as a promising mission for aerosol studies. This first results were presented to the 2001 EUMETSAT Meteorological Satellite Data Users’ Conference (Liberti et al., 2001).

Remote sensing of phytoplankton pigments at the basin scale
A total of 701 1-km remapped chlorophyll images covering the area (42-49N, 10W-42.3E) were available. Because one of the main objective of this task is to check whether the high chlorophyll values observed by satellite are artefacts attributable to the malfunction of the atmospheric correction algorithm, or are blooms due to atmospheric input of nutrients in the upper ocean due to the occurrence of dust events, we decided to develop an additional quality controlled chlorophyll products (level 4). Consequently, each chlorophyll map was flagged for clouds or other contamination factors using the corresponding 24 quality flag maps (McClain et al. 1995). This implies that case-2 water and possible contamination of chlorophyll by Saharan dust events an have been implicitly in the production of level 4 Mediterranean chlorophyll map time series. Then, daily composite and monthly chlorophyll maps were produced averaging all the available quality controlled data. On the basis of a preliminary list of dust events provided by Wp 1 we selected several dust events of characterized by different intensity to define a possible methodology to determine the a statistical relation between bloom occurrence and dust event. This analysis is ongoing. On the basis of a visual inspection of SeaWiFS true colour images, atmospheric optical thickness at 670 (AOT) and aerosol index products we select a region of 20x20 pixel around the area interested by maximum values of the aerosol index. A In this region time series of averaged value of chlorophyll was computed for a time window of 5 days before and 10 days after the event. In the same box an evaluation of the dust load was computed from the averaged value of AOT event by applying the Moulin et al (1997) relationship. In figure 3 an examples of this analysis is shown. The True colour images clearly show the occurrence of a dust event over the Western Mediterranean around the 22 of August 2000. On the basis of these images a box of 20x20 pixels in the Tyrrhenian Sea centred at (39.65N, 12.1E) was selected and the time series of the averaged chlorophyll content was computed. The aerosol index in the selected box has the value of 1.25 the 22 August. The dust load computed from SeaWiFS AOT corresponds to 0.35 mg/m2. In the lower panel of Figure 3 it is easy to recognize that an increase of 0.1 mg/m3 the chlorophyll occurs after five days from the event. This type picture often appears in any of the selected examples, but because of cloud occurrences and quality-flagged pixels not always the results are so clear. Moreover, on the basis of actual results, it is not easy to define an automatic method to define dust event and the area interested by the dust load that permits to avoid any visual and an easy analysis of the entire data time series.

References:
Baith, K., R. Lindsay, G. Fu, and C. R. McClain, “SeaDAS: Data Analysis System Developed for Ocean Color Satellite Sensors”, Eos, Trans. Am. Geophys. U., 82 (18), May 1,2001, p 202. O'Reilly, J. E., Maritorena, S., Siegel, D., O'Brien, M. C., Toole, D., Mitchell, B. G., Kahru, M., Chavez, F. P., Strutton, P., Cota, G., Hooker, S. B., McClain, C. R., Carder, K. L., Muller-Karger, F., Harding, L., Magnuson, A., Phinney, D., Moore, G. F., Aiken, J., Arrigo, K. R., Letelier, R., Culver, M. (2000). “Ocean color chlorophyll a algorithms for SeaWiFS, OC2, and OC4: Version 4”. SeaWiFS Postlaunch Technical Report Series, edited by Hooker, S. B and Firestone, E. R. Volume 11, SeaWiFS Postlaunch Calibration and Validation Analyses, Part 3. NASA, Goddard Space Flight Center, Greenbelt, Maryland. 9-23. Liberti G.L, F.D’Ortenzio, G.L.Volpe, R.Santoleri, M.Wang, C.McClain and V.E.Cachorro-Revilla, 2001. Validation of SeaWifs Aerosols products over the Mediteranean Basin. To appear in the Proceedings of the 2001 EUMETSAT Meteorological Satellite Data Users’ Conference. Antalya, Turkey: 1-5 October 2001. pp.8. McClain C.R., K. Arrigo, W.E. Esaias, M. Darzi, F.S Patt., R. H. Evans, J.W. Brown, C.W Brown, R.A. Barnes, and L. Kumar, SeaWiFS Algorithms, Part 1, NASA Tech. Memo. 104566, Vol. 28, edited by Hooker, S.B., Firestone E.R., and Acker, J.G., NASA Goddard Space Flight Center, Greenbelt, Maryland, 45 pp., 1995. Moulin, C. Guillard F., Dulac F. and Lambert C.E., 1997. Long-term daily monitoring of Saharan dust over ocean using Meteosat ISCCP-B2 data. 1. Methodology and preliminary results in the Mediterranean. Journal of Geophysical Research, 102: 16,947-16,958.


Figure 1: SeaWiFS monthly averaged maps of aerosol optical thickness (a), Angstrom coefficient (b) and chlorophyll concentration (c) for May 2000.

low qualityin situ data

 

Upper threshold of SeaWiFS aerosolmodels

Avignone
Oristano
Lampedusa
Venezia
Erdemli
Nea Michanionia
Sede Boker

Figure 2: Validation of the SeaWiFS atmospheric products: a) scatter plot of the aerosol optical thickness at 670 nm obtained from AERONET and SeaWiFS data b) scatter plot of Angstrom coefficient between 670 and 875 nm obtained from AERONET and SeaWiFS data. The different symbols refers to AERONET Stations. (c) Geographical location of the Mediterranean AERONET stations selected for the SeaiWiFS validation.


TIME (Julian days)
Chlorphyll-a
(mg/m3)


Figure 3: Sahara dust over the Western Mediterranean Sea from SeaWiFS data. Upper panels: True colour SeaWiFS images for 21, 22 and 23 August 2000. Lower panel: time series of the averaged chlorophyll content in a box 20x20 pixels centred at 39.65 N, 12.1 E in the Tyrrhenian Sea. The thick line at day 235 corresponds to 22/08/00 central day of the dust load event in the box.