Software: ESRI ArcView ver 3.1

Viola ver 1.0

Borland C++ Builder ver 4.0

Hardware: PC architecture, optimal with 2GB RAM

Operating System: 32-bit Microsoft Windows

Aim

Our project seeks to model larger fauna (mammals and birds) using an individual-based modelling approach. The essence of this approach is that an individual can, by virtue of its position in space or time, be influenced by its environment independently of other individuals. Each individual therefore has its own spatial position and is influenced by its own local environment. The approach is therefore spatially explicit and requires the use of spatial data.

The Simulation Framework

The simulation runs on a simulated landscape of 10 by 10 km (100 km2). The basic data for this landscape is stored in ArcView. This data is in the form of polygons representing features such as buildings, fields, forests and lakes in the landscape. These features are themselves dynamic, thus crops must grow in the fields and change the spatial structure of the field with time; farmers must use a crop rotation and roads must carry traffic. However, ArcView cannot deliver data fast enough to keep up with the demands of the animal models and it cannot be used to efficiently model the dynamic aspects of the features in the landscape. Therefore data must be transferred to the simulation program which can handle these aspects.

Data Transfer

This could be achieved by the development of complex algorithms for dealing with vector polygons and then writing these efficiently in a language such as C++. However, this requires a highly skilled programmer and will always suffer from the inefficiencies of position location when using polygons. The alternative, which has been adopted, is to save the data in a raster format and to access this directly from the simulation program.

Physical Data Collected

Data for the landscape is obtained from a number of sources. The main features are digitised from ortho-photos. Field boundaries are also digitised from ortho-photos, but the crop types are determined from Fødevareministeriets markblokkorte. Buildings are available from the T0 map and need only to be checked for changes since this data was gathered. Data resolution is 1 meter, however there is an effective minimum polygon size, such that long thin polygons like roadside verges are included, but a small patch of vegetation may be ignored. This is because the narrow linear features are often very important for animals.