Suitability of interpolation for plant community distribution varies depending on perennial versus annual species

Background/Question/Methods

Location of plant species in salt marshes have been suggested to depend mainly on the differing species tolerances to the varying abiotic factors.  At the Diamond Y Spring Preserve in west Texas, an inland salt marsh, spatial and seasonal fluctuations of several abiotic characteristics have been found to be highly variable and dependent on fluctuations in the local arid climate.  Plant studies have found several major perennial plant communities within the salt marsh including Sporobolus airoides, Schoenoplectus americanus, and Distichlis spicata as well as the federally listed annual Helianthus paradoxus sunflower.  The perennial communities are well delineated while the H. paradoxus population and distribution varies considerably on an annual basis.  The purpose of this study was to determine and map the abiotic factors of water table depth, soil pH and soil salinity in the salt marsh of the Diamond Y Spring Preserve through one growing season as well as estimating the plant cover of D. spicata, H. paradoxus, S. americanus, and S. airoides.  Geographical Information System (GIS) software was then used to investigate plant distribution based on specific abiotic environmental characteristics at a landscape or community scale. 

Results/Conclusions

Due sparse vegetation on higher elevations combined with the clumping bunchgrass habit of the Sporobolus airoides the mean plant cover in the salt marsh was 59%.  From driest to wettest, the sequence was a Sporobolus airoides grassland (26% of marsh), a Distichlis spicata grassland (50% of marsh), and in the drainage a Schoenoplectus americanus marsh community (2% of the marsh). An Inverse Distance Weighted interpolation created a spotty surface with data artifacts and moderate success at predicting the perennial distribution.  Conversely using an Ordinary Kriging technique resulted in a smoother prediction surface which was more successful at predicting the perennial distribution.  The opposite was true for predicting the H. paradoxus distribution.  Using Ordinary Kriging no relationship was found between the abiotic characteristics and the annual’s distribution while a specific relationship was found using Inverse Distance Weighted interpolation.  Locations in the marsh with H. paradoxus were found to coincide only with locations with depth to water greater than 25 cm, soil salinity between 7 and 12 g/kg, and soil pH between 8.3 and 8.5.