GISdevelopment.net ---> AARS ---> ACRS 1989 ---> Oceanography

Determination of primary production in the Japan coast the data of NIMBUS-7,CZCS

Yasuhiro Sugimori, Hajime Fukushima, Yasuhiro Senga
School of Marine Science and Technology, Tokai University
Orido, Shimizu, Japan


Abstract
The phytoplankton pigment concentrations around Japan inland sea were observed using data of Coastal Zone Scanner (CZCS). Distinct meso-scal,e eddies at Oyashio-Khuroshio front and detailed mixing patterns could be inferred sixth the pigment picture. Highly reflective patterns were observed in each channel of image in late spring.

Through image processing, it is found that aerosol types may be different in the Pacific side and the Japan Sea, where the Godo-Clark scheme will not be applicable. The in-water algorithm might also be different.

Introduction
A typical example of the application of the NIMBUS-7, CZCS was the study of warm core ring structure (Gordon, 1982). The CZCS was the best satellite sensor for the determination of distribution of phytoplakton pigment. As fundamental study to formulation of a chlorophyll algorithm Hovis (1980) examined the spectral patterns of upward irradiant from water of various chlorophyll concentration (in the paper, Chlorophyll or pigment equal to mean Chlorophyll plus Chaeophytin). Fig.1 shows clearly that the upward irradiance patterns vary strongly with chlorophyll concentration. Thus, the distribution might be easily derived from data of each channel. The upper indication in Fig. 1 shows the band width of each channel of which CH-4 can present the definite absorption of light by chlorophyll production.

Model for satellite determination of chlorophyll
It is well estimated by all optical oceanographers that Gordon-Clark Model, using the optical-nol method in near infrared radiation emitted from ocean surface, can determine the distribution of chlorophyll in the upper ocean. Fig 2 shows their regression lines (dashed lines). The relationship between chlorophyll content and spectral irradiance varies because of changes with season, religion and species. This is shown in Fig.2 by data from S&W refer to Morel, Clark Gordon, Sugihara and smith & Wilson respectively

The solid line in Fig.2 is the regression of our data from around Japan coast. Thus, the chlorophyll content can be found through the spectral irradiance by,

Log C = Log a + b Log R ----------------------(1)

R =Ch-1/Ch-2m or Ch-2/Ch-3 ---------------------------(2)

Unfortunately, our field data were not coincident with the CZCS images, so we used Gordon-Algorithm as the first attempt.

Determination of chlorophyll distribution
Picture 1 shows the chlorophyll l distribution at the south eastern end of Main-island of Japan. The Euroshio, which has low chlorophyll concentration, is flowing from SW to E. The Oyashio and coastal water with high chlorophyll concentrations are trapped by large frontal eddy. Even south of the Kusoshio, a high chlorophyll concentration can be distinctly seen. Thus, there must be eddies causing upwelling and vertical mixing to support the primary production.

Fig.1 Upward spectral irradiance just below the surface by each ch.1 content water
(Hovis, 1990)


Picture 1, CZCS-derived pigment image(1981, April,14)

Picture 2 presents the pigment distributions off south Hokkaido. Oyashio water flows from the north with high chlorophyll coastal water. After detaching from Hokkaido coast to the south, this water mass will collide with Tsugaru water core. The ocean condition related to collision can be seen at the strong front south of Hokkaido. The Oyashio water forms large eddies as it flows toward the south. The temperature distribution is almost the same as pigment one for this case.

Fig. 2 Regresion line from each observers and data around Japan.


Picture 2, CZCS-derived pigment image
(1981, Sept.,17)

Picture 3 shows pattern of typical perturbed area where is located in the inland sea of Japan Pacific coast. It is clearly recognized that the higher concentration of chlorophyll distribution is accumulated in the area crowded by the small islands and in the non-water Mass exchanged area.

Picture 3, CZCS-drived pigment image
(1982, May,23)

Comparison with ship-sampled measurement
CZCS-derived pigment concentration in three images were compared to JMA ship-sampled data sets /4/, which were obtained within 5 days from satellite 0verrpasses. In fig. 2 (a), satellite-derived concentrations were calculated from LW(520)/LW(550), since corrected LW(443) can be negative, causing discrepancies from the sea-truth. Fig. 2 (b) shows fairly good agreement except for high pigment concentration cases. This disagreement might be explained by the difference in sampling method or by limitation of the estimation algorithm. The comparison in Fig. 2(c), which has relatively low pigment range, also shows good agreement except for the data very low concentration

To summarize, the satellite-derived estimates seemed to be in good agreement with the sea-truth which has the pigment range of 0.1 to 5.0 g/1, although some deviation may arise for other cases.

Conclusions
Since CZCS images give chlorophyll distribution that interpret ship observational data well, we consider that, to a certain extent, Gordon-Clark approach is workable for the data around Japan. But for more precise estimation, the following aspects should be noted.

  1. As pointed out, "clear water area" that serve as an anchor point for the atmospheric correction, is not necessarily present in every scene.

  2. There are high sediment loading in the East China Sea and the Yellow Sea, where the Gordon-Clark scheme will not be applicable. The in-water algorithm might also be different.

  3. Aerosol types may be different in the Pacific side and the Japan Sea side. As an example, the effect of "Yellow Sand" transported by winds from the main land China, Which is often observed in March-May, might be considered.


    Fig.2,Comarison between ship-observed(solid points) and CZCS-estimated(solid line) surface pigment concentrations. Attached figure with each point indicates number of days past the satellite overpass.

  4. Estimated water-leaving radiance at 443 mm often becomes negative, suggesting the necessity of more accurate evaluation of Raleigh scattering path radiance in atmosphere.
References
  1. Gordon H., D. Clark, J. Brown, R. Evans, and W. Broenkow (1983) : phytoplankton pigment concentration in the middle Atlantic Bight : comparison of ship determination CZCS estimates, applied optics, 22,1,20.

  2. Gordon H. and A. Morel (1983) : Remote assessment of ocean color for interpretation of satellite visible imagery. Spring-Verlag.

  3. Holligan P., V. Harbor P. Camus, and M. Champagne-Philoppe (1983) : Satellite and should studies of cocolithophore production along a continental shelf edge, Nature, 304, 28, 339.

  4. The Japan Meteorological Agency, The results of marine meteorological and oceanograd observations, 67, 69 and 70.

  5. Ogishima T., H. Fukushima and Y. Sugimori (1986) : problems relating to atmospheric correction algorithm for CZCS data in Japanese coastal area, Bull. Airborne and satellite phys. & Fish. Oceanogr. 8, 53.