Ten thousand scale vegetation maps are necessary to evaluate and analyze the urban environment, since vegetation can affect urban environment on thermal aspects. As a method for constructing the vegetation cover map, the vegetation cover ratio in a pixel using the index is thought as convenient method for estimating the small scale vegetation in urban areas. This study aimed to find out an index in order to assume the vegetation cover ratio in a pixel in urban areas.
The relationships between vegetation cover ratio in a pixel and conventional vegetation indices, NDVI, RVI and SRVI, were numerically analyzed. The relationships between those were analyzed by the material combina-tions of the simple urban surfaces, which have vegetations, tree and lawn, and non-vegetations, concrete, soil, asphalt, roof tile, aluminium and tile. The results shown the possibility that the estimation accuracy of vegetation cover ratio in a pixel may be unbalanced according as amount of the vegetation cover ratio in a pixel and as characteristics of the site. On that reason, we proposed a new index, modified ratio vegetation index (MRVI), which have three concepts of index for extract the vegetation cover ratio in a pixel in urban areas, adaptable to general satellite data, approximate by linear calibration, small assumption error. The efficiency of MRVI was proved by the material combinations of the simple urban surfaces and also simulation MSS data with several spatial resolutions.

View full abstract

The new vegetation index was developed by the pattern decomposition method taking into account the requirements that the index should be linear to the amount of leaves and the vegetation vigor. To validate this vegetation index, its relations to the thickness, chlorophyll content and water content of leaves were studied using spectral reflectance measurements and compared to NDVI. The results clearly show that this vegetation index obtained by the pattern decomposition method is sensitive to the amount of leaves and to vegetation vigor, and satisfies the requirements for an ideal vegetation index more comprehensively than NDVI.

View full abstract

Artificial oil slick observations with a real-aperture Side-Looking Airborne Radar (SLAR) were conducted on Nov. 10, 1991 and Nov. 1, 1992 at 100 km off the coast of Japan in the Pacific Ocean. The SLAR obtained X-band VV-polarized radar images of an oil slick in synchronization with the synthetic aperture radar of European Remote Sensing Satellite 1 (ERS-1 SAR). The oil slicks were created by spilling oleyl alcohol from a small ship. Sea truth data were collected by a research vessel and the small ship. The SLAR has a 40-degree fanbeam in the cross-track direction. The sea surface backscatter depression, called "damping", caused by the oil slick was measured at various incidence angles from 20 degrees to 58 degrees by SLAR. The sea surface wind speed was 11.3 m/s in the experiment of 1991 and 6.6 m/s in the experiment of 1992. The magnitude of damping was plotted as a function of the incidence angle. There are two notable features in the plots. The first one is that the damping in 1991 is smaller than that in 1992 at all incidence angles, because of the difference in wind speed. This tendency is consistent with the results of other oil slick experiments in which lower wind speeds cause larger damping. The second feature is that there is no clear peak in the plots. The value of damping increases with incidence angles up to 40 degrees, but it is not clear whether it further increases with incidence angles beyond 40 degrees.
While other researchers often found a peak in the damping at a certain incidence angle around 50 degrees, the existence of such a peak is not clear from the data obtained in our experiments. The SLAR data are compared with ERS-1 SAR data taken simultaneously. The incidence angle of ERS-1 SAR is fixed at 23 degrees. The damping values obtained from these measurements are in good agreement on both days. In addition, the amount of damping obtained is well coincident with those estimated from the airborne scatterometer measurement.

View full abstract

A method for Sea Surface Temperature (SST) estimation with Linearized Inversion of Radiative Transfer Code is proposed. 23 to 32% of improvement of SST estimation accuracy was confirmed with a variety of the atmospheric parameters compared to the existing method of Split Window. The improvement depends on the Ocean Areas. The best improvement (56%) was occured for SubArctic Winter while the worst improvement (7%) was occured for Tropic. For all of the ocean areas, more than 30% of improvement was achived with the proposed method. By using the most adequate parameters for each the Ocean Areas such as Tropic, Midlatitude, SubArctic and for each seasons, summer and winter, 4% of improvement was achieved compared to the method with only one set of parameters.
The accuracy of our proposed Linearized Inversion for SST estimation with Thermal Infrared data was evaluated with the real ADEOS/OCTS data. It was 1.1 (K) and was not so poor if a skin and bulk temperature difference is taken into account.

View full abstract

We investigated the relation between the water quality and Landsat TM data observed in Tokyo bay from 1986 to 1992, and the following results were obtained.
1) A high correlation between surface water temperature and Band 6 was found.
2) It was also shown that there was a high correlation between observed water quality (chlorophyll-a, transparency) and calculated water quality with Band 1 and Band 4 of TM data.
3) Images of surface water quality were obtained succesfully by using the relation of observed data with Landsat TM data.

View full abstract