NDVI from satellite data offers various kinds of physical values related with plants. Landsat TM has high resolution and gives the seasonal change of NDVI for each land cover. Then, we proposed seasonal change models for NDVI using a logistic curve. NDVI for each land cover except water surface fitted very well with logistic curves for less than 300 Julian day. We modified the logistic curve adding a negative-slope line. This modified logistic curve expresses periodic change and gave good fit with NDVI. On the other hand, leaf area, leaf length, dry weight, and spectrum reflectance were measured for paddy pots in greenhouse cultivation. These results led to a series of relationship among these parameters. Finally, the relationship between NDVI and leaf area index was obtained. Using these regression curves for NDVI, the temporal and spatial distributions of LAI of paddy fields in the watershed could be estimated.
Quantitative evaluation of the surface roughness is a key to detect the surface soil moisture from the active microwave remote sensing. This study develops a new inversion algorithm of the surface roughness from a viewpoint of the hydrological application. Field experiments were conducted using C-band scatterometer system mounted on a car and the backscatter from the bare soil surface was observed with multiple angles at the same time. Measured surface roughness parameters vary and are very difficult to determine the representative roughness parameters even in a single field. The inversely obtained ones with the assumption that the soil moisture is known, does not agree with the measured ones, but the dependency of the backscattering coefficients computed from the theoretical model with the inversely obtained parameters, on both the incident angle and the soil moisture, shows good agreement with the observed ones. The result can be explained theoretically and terminology "surface roughness factor" is proposed to evaluate the surface roughnesses quantitatively.
Direct solar radiation measurements were carried out on board in the Osaka Bay, on March 18, 1997, and at Sand Dune Agriculture Experiment Station, Ishikawa, on April 26, 1997. Based on those data, aerosol optical thickness at 550nm (τa550) and Angstrom exponent (α) were estimated. They were also retrieved from the degree of polarization at wavelengths of 443, 670, and 865 nm measured quasi-simultaneously by the POLDER sensor onboard the ADEOS satellite. The τa550 values retrieved from the degree of polarization were compared with those from the direct solar radiation measurements. The difference was about 0.03. The α values estimated from the degree of polarization were smaller than those from the direct solar radiation by 0.5 in typical cases.
A method for cross calibration of different sensors onboard same satellites is proposed. BRDF of the standard plaque for the reference of the surface BRDF measurement is measured and corrected in the proposed cross calibration. Furthermore, difference of the spectral responses of the sensors as well as differnt Instantaneous Field of View: IFOV are taken into acount in the cross calibration. As the results from the experiment with ADEOS/AVNIR and OCTS, it is found that the proposed method shows a good cross calibration between both and works for evaluation of NEdR or Signal to Noise ratio: S/N. Through a comparison of the evaluated S/N with the previously reported S/N from NASDA, it is also found that the proposed method is appropriate for the S/N validation.