Journal of Uncertain Systems

Ecological data is very costly and difficult to collect, and quite often the sampled data are insufficient for further spatial analysis. Today, we as spatial modellers are often presented with the situation whereby a set of data is collected already, although from the viewpoint of spatial analysis the data is insufficient, but re-sampling is impossible because of the cost and time limits. In this paper, we are dealing with a two spatial problems whereby: the data is just species presence only and no numerical data; and also the data sampling is not well spread over the study area. These are two very common problems that spatial modellers face everyday, and in this paper we provide some simple techniques to deal with these problems. We firstly use frequency counts to deal with species presence data, then use the recently developed partial differential equation motivated regression (PDEMR) model to predict the unknown locations, and finally combine these data to produce a kriging prediction map. These techniques are fairly new, but very effective in dealing with ecological data problems. For illustration, Protea rare species ( i.e., the population size of 10 to 100), in the Cape Floristic Region, from 1992 to 2002, South Africa, are used.