A GIS-based plant prediction model for wetland ecosystems

摘要: An existing non-spatial model for the prediction of response wetland plant species on ecological factors has been transformed into a GIS-based which produces spatial output at landscape scale. The input, patterns factors, were constructed with geostatistical interpolation (kriging). With this presence and absence 78 are predicted an area wetlands in Netherlands approximately 500 square kilometers. validated, estimated uncertainty input propagated through model. At level shows spatially coherent non-random patterns. validation is affected by propagation errors number valid predictions declines 10–20% when 95% confidence intervals used validation. This study that it feasible to use method construct scale these as added values however, affects