| We developed individual-based models (IBMs) to assess the effects of changes in the physical environment due to river improvement works on aquatic animals that use habitat networks composed of rivers and agricultural channels. As an appropriate indicator species for assessing habitat networks， we focused on the migrating freshwater shrimp Caridina leucosticta， an amphidromous species inhabiting the rivers and agricultural channels of Japan. We selected a research site in an agricultural area in central Japan and precisely modeled the physical environments and factors affecting both shrimp migration into agricultural channels from a river at the site and shrimp population viability. We modeled the topography of the river and the agricultural channel network (modeled as the virtual water area， VWA); the current velocity and water depth; the height difference between the river and agricultural channels; and the presence of water， and water temperature， in the agricultural channels. In addition， we defined a migration index for assessing suitability for shrimp migration between the river and each channel . In modeling the life history stages of C. leucosticta， we defined a “virtual shrimp” class and developed a program covering the life history of this species in detail. We used this program to model the virtual shrimp’s migration into， and growth in， the VWA in accordance with shrimp spatial preferences and life history. After the release of many virtual shrimps into the VWA， we simulated and predicted their distribution patterns. To verify the model’s reliability， we conducted field surveys of the distribution of the shrimps in the river and agricultural channel networks at the research site during 2008 and 2009， and we compared the distribution in the field survey with that in the simulation. The distribution pattern simulated by the IBMs was similar to the field distribution， indicating that use of the IBMs was a valid method for evaluating habitat networks. To simulate the distributional patterns of this shrimp more accurately， we need to improve our model by focusing especially on the migration index and shrimp behavior. |
