Developing a wetland condition prediction model using landscape structure variability

Abstract

Recent studies found substantial variability in plant community integrity of wetlands in the Prairie Pothole Region (PPR) of central North Dakota, USA. We speculated that this variability might be connected to the nature of the surrounding landscapes and that a link might exist between landscape spatial metrics and wetland condition. We explored this potential link, using a case study in the PPR. A combination of remote sensing, geographic information systems (GIS), and landscape spatial metrics was used to: 1) examine the condition-landscape pattern relationship of temporary and seasonal wetlands, and 2) develop a landscape-level decision support tool for rapid assessment of wetland condition. We sampled 73 wetlands in the study area. We used the Index of Plant Community Integrity (IPCI) as our measure of wetland condition. A wetland landscape was defined by a 300 m radius circular area (0.283 km2) around each habitat. Quantitative characterization of landscape pattern was conducted using metrics computed from land cover categorization maps processed from multi-temporal Landsat satellite data. Ordination of wetland samples in a multivariate space of landscape metrics using non-metric multidimensional scaling revealed strong associations between wetland condition and 10 landscape metrics, primarily among seasonal wetlands. The Landscape Wetland Condition Analysis Model (LWCAM) was developed and validated for rapid quantitative assessment of wetland condition. The model was based on three landscape metrics considered most important for use in the PPR: 1) grassland percent core area of landscape, 2) grassland largest patch index, and 3) the number of patches per unit area. We concluded that surrounding natural grasslands and landscape fragmentation were the most important influences on the structure and plant community condition of wetland ecosystems.