| Project/Area Number |
17580215
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Irrigation, drainage and rural engineering/Rural planning
|
|---|
| Research Institution | KYUSHU UNIVERCITY |
|---|
Principal Investigator |
HIRAMATSU Kazuaki Kyushu University, Faculty of Agriculture, Professor, 大学院農学研究院, 教授 (10199094)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
MORI Makito Kyushu University, Faculty of Agriculture, Assistant Professor, 大学院農学研究院, 助手 (60325496)
|
|---|
| Project Period (FY) |
2005 – 2006
|
| Project Status |
Completed (Fiscal Year 2006)
|
| Budget Amount *help |
¥3,600,000 (Direct Cost: ¥3,600,000)
Fiscal Year 2006: ¥500,000 (Direct Cost: ¥500,000)
Fiscal Year 2005: ¥3,100,000 (Direct Cost: ¥3,100,000)
|
|---|
| Keywords | Ecosystem concervation / Japanese medaka / Habitat preference / Agricultural canal / Artificial intelligence technique / Fuzzy reasoning / Genetic algorithm / Artificial neural network model |
|---|
| Research Abstract |
The fuzzy-rule-based mathematical models for quantifying the preference intensity of Japanese Medaka to the three environmental factors of water depth, current velocity and cover ratio were firstly constructed in laboratory open-channel experiments. In this model, a simplified fuzzy reasoning is introduced to explicitly take essential vagueness of fish behavior into consideration and a genetic algorithm was introduced to search for the optimal functional representation of preference intensity. The models were then verified by laboratory water-tank experiments and field observations. The results indicated that the fuzzy preference intensity model had a good performance in predicting the habitat preference and that the environmental factor of current velocity obviously affected the environmental preference of Japanese Medaka.
Next, the preference of Japanese Medaka for vegetation was evaluated using the fuzzy preference intensity model with interactions between water depth, current veloci
… More
ty, and cover ratio in an agricultural canal. For the objective to break through the difficulty in model construction with insufficient data observed in an agricultural canal, the model was conjugated with a model developed in laboratory experiment. And then, obtained model was assessed using AIC (Akaike's Information Criterion) in order to evaluate the significance of aquatic vegetation with a statistical approach. The habitat prediction with consideration of vegetation factor showed good agreements between predicted and observed fish population density. AIC score for the model with vegetation factor was markedly reduced compared to the model without it. These results suggest the significance of aquatic vegetation to the habitat selection of Japanese Medaka in agricultural canal.
Finally, artificial neural networks were utilized to predict the habitat selection of Japanese Medaka in agricultural canals. The fuzzy neural network model precisely predicted the habitat preference of Japanese Medaka in an agricultural canal, and the results showed a good agreement between calculated and observed habitat suitability indices. Less
|
