PotNat, een GIS-tool voor het bepalen van de abiotische kansrijkdom van natuurtypen: deel 1: methodologie

An adequate / a correct estimate of the (natural) potential for the occurrence of (pre-specified) nature types is a key component in planning processes which impact the spatial distribution and occurrence of nature types (vegetation types, habitat types, …). For this purpose, a number of predictive models have been developed. Some of them give accurate predictions, but do this only for certain nature types, have high demands on data or the use of complex mathematics makes the results difficult to interprete, others are more generalising but disregard local or up to date information. The need for a model that enables the translation of existing ecological knowledge throughout Flanders and on the scale of Flanders still exists. For this purpose, the Research Institute for Nature and Forest (INBO) has developed PotNat, a deterministic model based on the Dutch model Natles.

This report outlines the theoretical framework of the PotNat model. The model is based on two sources of a priori available knowledge. One source, called the ‘demand side’, encompasses the abiotic requirements (site conditions) of a given nature type. The other source, the ‘supply side’, provides the physical environment at the locations (of interest / under study). Both knowledge sources are described using random variables. Hence, the abiotic potential of a location for a given nature type is defined as the product of two probabilities: the probability that the nature type can prosper under a specific environment times the probability that the particular environment occurs on that location. When applied in practice, some simplifications have been applied tot his concept. For instance, not for all unique combinations of the physical variables it is known whether they are part of the site conditions. Moreover, checking all possible combinations would need (too) much computing effort.
Instead, the abiotic potential is calculated in two, sometimes three steps. In the first step, the extent to which a particular location meets the requirements of a nature type is calculated separately for each selected physical variable (which was transformed into a class variable). In the next step, results are converted into a global score by means of a simple multiplication or a weighted average. If for a nature type a location is not suitable for any of the physical variables, no abiotic potential is allotted. Particular attention is paid to how missing local environmental data are dealt with. Through the applied simplifications the end result must not be seen as a real chance, but rather as a score that needs to be calibrated with field observations.

In the second part of the report, we describe also an application of the model to use in and on the scale of Flanders: PotNat Flanders. The nature typology is based on a specifically designed hierarchical system of ecotope and ecoseries types well as the Natura 2000 habitat typology. The demand and supply sides are described by a fixed set of nine physical variables: soil profile, soil texture, lowest groundwater level, flooding tolerance, soil nutrient content, water chemistry, water regime, water salinity and soil acidity. Using these nine variables, a classification has been designed taking into account the ecological requirements of the nature types as well as the spatial information available for Flanders. The site condition requirements have been described based on current ecological knowledge. The spatial information was mainly based on some area-wide information sources such as the biological valuation map and the soil map. The database ‘Aardewerk’ and flood maps gave additional information. A detailed description of the methodology used to generate GIS layers is provided. To calculate the nature potential a tool for ArcGIS was designed.