Direct Specification of Distributions

Within the pipeline for the production of geometric data, also the plant distri­bution is divided into substeps. in a first step, areals are specified and attributed with the characteristics of their vegetation. This step is usually implemented using a geographic information system (GiS). Due to an uniform geodetic ref­erence system, this type of system allows the storing and processing of land­scape data independently of its physical resolution.

GIS ^ In a GIS geographical data of different logical levels is arranged using a refer­ence coordinate system. In parallel to the digital terrain models, in which the geometry of the landscape – for example, a synthetically produced elevation model – is stored, aerial views can be included, and measured values, single objects as well as many other data types can be called up for visualization and simulation.

All geometrical objects, such as roads, houses or individual regions (stands) here are assigned attributes. These can be used for retrieval queries or visual­ization. For the production of plant distributions the stands of the plant com­munities occurring in each case are indicated as attributes.

The final result of this rather abstract modeling is therefore a file with a set of polygonal descriptions of individual stands, in which only one plant commu­nity occurs in each case. Each individual stand is filled in the next step with plants. This procedure corresponds to the module in Fig. 8.1 that generates the ecosystem file with the list of the plants as the final product.

As already mentioned in the introduction, there are different options for these techniques. In the following we differentiate between explicitly given plant distributions and those that result from generation procedures. The production of a complex ecological system is relatively tedious; it allows however, deal­ing in detail with constraints as well as local characteristics. Therefore, it is a preferred methodology with smaller or intricately covered surfaces.

Explicitly given distributions are advantageous if large areas must be filled or if human intervention modifies the natural development of ecological systems. Often also a combination of both techniques is meaningful, such as when larger elements on a plane (e. g. trees in a meadow) are explicitly to be modeled and the remaining plants have to be grouped in between them.

An explicitly given plant distribution may, for example, consist of an analysis aerial photographs ^ of satellite images or mappings produced by aerial photographs of real existing ecological systems. The satellite Landsat supplies images in several channels from the infrared to the ultraviolet range. Since many species possess char­acteristic reflection characteristics in different frequency ranges, their position

in a plant community can be obtained from the evaluation of these reflection characteristics.

In a mapping, positions and plant types as well as further parameters are stored as, for instance, the growth rate. The resulting distributions can be used as starting points for areas to be filled in a similar way. For example, they can be used in particular in places where uniform distributions are required that are not bound to special local conditions.