Plant Ecology

, Volume 169, Issue 2, pp 215–226

Spatial distribution and performance of Scirpus americanus ramets across a temperate intertidal marsh resource gradient

  • Jim D. Karagatzides
  • Harry R. Manson
  • Leonard J.S. Tsuji

DOI: 10.1023/A:1026015218282

Cite this article as:
Karagatzides, J.D., Manson, H.R. & Tsuji, L.J. Plant Ecology (2002) 169: 215. doi:10.1023/A:1026015218282


Intertidal wetlands provide examples of plants whose performance varies in response to resource (elevation) gradients, such as light, that are primarily determined by differences in tidal flooding. The spatial distribution (using Thiessen polygons to delineate neighborhoods) and performance (ramet height, weight, basal perimeter, photosynthetic area and inflorescence weight) of Scirpus americanus Pers. ramets were measured for three plots in each of the high and low elevation extremes of a resource gradient on an intertidal marsh of the Fraser River Delta, British Columbia, Canada. Most neighborhood polygons in both sites were six-sided and their frequency exceeded that expected for a Poisson distribution (i.e., complete spatial randomness). The frequency distributions of polygon shapes between low and high marsh sites were not significantly different but there were more five-sided polygons than expected in the high marsh. Two-way ANOVA for polygon area revealed significant differences between sites (p < 0.001, high marsh polygons ∼40% the area of those in the low marsh and with significantly less variability) and shape (p < 0.001, area increases with increasing number of neighbors) but the interaction effect was not significant (p= 0.10). There were no significant correlations (p > 0.05) between polygon area and ramet performance except for one high marsh plot that had significant (p < 0.01) positive correlations for polygon area and ramet photosynthetic surface area, weight and basal perimeter. Moran’s I and Geary’s C indicated that there is generally no spatial autocorrelation for polygon shape or ramet performance as only one plot in the high marsh had a significant (p= 0.009) negative autocorrelation for ramet height. Nearest neighbor analysis revealed a clumped distribution of ramets for two low marsh and one high marsh plot; ramets in the three remaining plots were distributed randomly. Ripley’s K revealed that ramet distributions were evenly dispersed at spatial scales of ∼2 cm for one plot in the high marsh and at ∼12 cm for one plot in the low marsh. We conclude that the overall lack of an even distribution across this resource gradient may be due to a heterogeneous environment at a scale of centimeters, asynchronous shoot emergence throughout the growing season, and disturbance by grubbing geese.

Clonal plant Nearest-neighbor analysis Ripley's K Spatial autocorrelation Spatial distribution Thiessen polygons 

Copyright information

© Kluwer Academic Publishers 2003

Authors and Affiliations

  • Jim D. Karagatzides
    • 1
  • Harry R. Manson
    • 2
  • Leonard J.S. Tsuji
    • 3
  1. 1.Department of GeographyQueen's UniversityKingstonCanada
  2. 2.Environmental Applied Science and Management ProgramRyerson UniversityTorontoCanada
  3. 3.Department of Environment and Resource StudiesUniversity of WaterlooWaterlooCanada

Personalised recommendations