In FY 1996, we published 18 papers concerning remote sensing methods for monitoring a variety of crops, isotope analyzing methods, statistical methods and their applications, neural network model analysis of plant shape data, and other related topics. Mr. Mino's article "Monitoring Annual Changes in Grasslands using Multi Temporal Satellite Data" received a young scientist's award from the Japan Society of Photogrammetry and Remote Sensing. Dr. Mei and Dr.Yamasaki applied to the Japan Patent Office for their invention on the "Interface for Coupling Capillary Electrophoresis to Inductively Coupled Plasma Spectrometry". Twelve short papers were published in 6 journals. Twenty nine themes were presented in 16 symposia of different disciplines. Fifty scientific or tutorial writings were published on various media. The Seminars of the Information Analysis Division were carried out 9 times. Six foreign visiting scientists and 11 national researchers stayed in our division.
Topic1
Transfer Factors of Radionuclides from Paddy Soils to Polished Rice
Polished rice occupies an important place as the country's staple food. Therefore, the determination of various radioactive fallout nuclides in polished rice contaminated by nuclear tests, and the estimation of transfer factors of these nuclides from the soil to the rice,are important for providing basic information for evaluating the risks of human ingestion of these nuclides and explaining the physical and genetic effects of low doses of radiation on the human body.
The present studies were conducted in order to estimate the transfer factors of radionuclides from paddy soil to polished rice using the 90Sr and 137Cs survey data and analytical data obtained by conventional radiometric and ICP-MS methods. Soils and rice grown on these soils were collected nationwide from 15 National and Prefectural Agricultural Experiment Stations in 1990.
The Figure shows the transfer factors (element concentrations in polished rice divided by elements concentrations in soil) for the 16 elements (19 nuclides) calculated from the total analyses data for paddy soils and polished rice grown on these soils. The highest transfer factors were found in the alkali metals, followed by the alkaline earth metals. Last came the lanthanoids (Eu > La, Ce, Nd, Sm, Gd, Lu), all of which (excluding Eu) showed extremely low values. The transfer factor (×10-3) for 90Sr was 4.8 ± 3.4 and that for 137Cs was 2.6 ± 2.8. While they were roughly equal, 90Sr tended to be slightly higher than 137Cs. A comparison of the transfer factors showed that radioactive Sr (fallout) was 5.8 ± 3.8 times higher than natural Sr, while radioactive 137Cs (fallout) was 3.8 ± 2.2 times higher than natural Cs. These differences between natural and fallout nuclides can be attributed to the different chemical forms of these nuclides in soils. (Fig.1)
Topic2
Linking Remote Sensing and Growth Simulation for Predicting Potential and Actual Crop Growth and Yield
A computer-aided modeling approach has significant potentials in agriculture such as for the prediction, diagnosis and management of plant and environmental dynamics. However, it is not easy to take into account all plant and environmental factors in a model such as water, nutrients, soil, disease and insects. Complicated mechanistic models usually require too many parameters and/or coefficients as inputs. It is tedious or sometimes impossible to collect all input data and, even with all inputs given, the improvement of the accuracy may be questionable. Thus,a within-season parameterization is an effective approach to reduce the model complexity, to simplify input requirements, and to make the models more easily used for operational purposes. We have proposed a simple approach for monitoring and predicting the actual and potential crop growth and yield.The approach is based on the combination of a crop simulation model and remotely sensed spectral measurements over plant canopies.The combination of the crop growth model and remote sensing was realized by an automated re-calibration module which was designed to utilize remote sensing data for producing an optimum set of ecophysiological parameters such as leaf area index and radiation use efficiency in the simulation model (Fig.2).
In a case study,a simplified process model was used which simulates the daily growth and yield of irrigated rice based on daily weather data. Spectral measurements were obtained by a portable radiometer with seven bands from the visible to mid-infrared wavelength region.Plant parameters such as leaf area index were well estimated from spectral data, and used as inputs to the re-calibration module.
The leaf area index (LAI) was used as a key factor for checking the model accuracy. The case study showed that the combination of remote sensing and crop growth models has great potential for the accurate and timely prediction of plant growth and yield (Fig.3).
This approach also proved useful for estimating realistic values of ecophysiological parameters,such as the maximum leaf area index and overall radiation use efficiency in individual canopies, from remote sensing measurements made during the season.
Fig.1 Distribution of transfer factors from soil to polished rice
* These elements include natural radioactive nuclides
Fig.2 Scheme of the real-time calibration system
Fig.3 Parameterization and prediction at the middle growth stage based on the real-time calibration system with all available remote sensing data.
Meas, Sim, and Sim+RS[#] mean measured data, simulation results and results of simulation + remote sensing, respectively. The # indicates number of remotely sensed data used. Estimated values of parameters are as follows;
| DVIi | LAIi | DWi | LAIn | CE | |
|---|---|---|---|---|---|
| Sim: | 0.250 | 0.200 | 20.00 | 5.50 | 1.95 |
| +Rs: | 0.133 | 0.039 | 13.54 | 3.42 | 1.79 |
where the DVIi, LAIi, DWi are initial values of the developmental index, LAI, and dry matter, respectively, and the LAIn is the asymptotic value of leaf area growth, and the CE is radiation use efficiency.