Loading ...
Sorry, an error occurred while loading the content.

1503

Expand Messages
  • Juliann Aukema
    Feb 29, 2000
    • 0 Attachment
      Thank you all for your suggestions. Here is a summary of responses
      following my original question.

      > Hello.
      >
      > I've been watching this list for only a short time and I hope someone
      > out there may be able to help me. I'll try to be brief.
      > I am looking at patterns of parasitic plant infection on tree hosts. I
      > would like to address questions such as: Are infected individuals
      > clustered? If so, at what scale? Is the clustering related to host
      > distribution or size? to underlying habitat heterogeneity? What is the PDF
      > of the seed distribution with respect to the parents? Are seed deposition
      > rates a good proximate measure of the infection rates? I am examining
      > processes and mechanisms of parasite transmission in other experiments and
      > observations of vector behavior. And I hope to integrate both parts into a
      > model of parasite transmission and spread.
      > I have mapped (ARCVIEW) trees, parasites and seeds (infective
      > propagules) on approximately 4 hectares (about 900 trees) using GPS. I'm
      > not sure this is a large enough scale at which to detect patterns. I am
      > wondering if there is a way to further subsample at a larger scale
      > (complete mapping at a larger scale is too time-consuming right now). Given
      > that I need to sample several non-contiguous areas for another part of the
      > project (before my field season ends and before I will be able to analyze
      > what I already have), perhaps I could choose locations that I could use for
      > the spatial analysis as well. For example could I use the plot I have
      > mapped and sample at different lag distances? Would I have to sample in
      > concentric circles or could I just sample points? What would be a point? 50
      > trees? 1 tree? in a circle? line?
      > Thank you very much. I would be grateful for any suggestions.

      I was directed to the following websites and articles:

      * POINT PATTERN ANALYSIS at:
      http://xerxes.sph.umich.edu:2000/cgi-bin/cgi-tcl-examples/generic/ppa/ppa.cgi

      * you might want to look at some of our on-line modules, especially
      http://xerxes.sph.umich.edu:2000/geomed/modules/cluster/index.html

      * to download Rookcase:
      http://aix1.uottawa.ca/academic/arts/geographie/lpcweb/sections1/software/fr
      msoft.htm
      *
      http://ag.arizona.edu/PLP/GIS/

      * Oliver and Webster 1986. Combining nested and linear sampling for
      determining the scale and form of spatial variation of regionalized
      variables. Geograph. Anal. 18:227.242.

      * Sokal and Oden 1978 Spatial autocorrelation in biology 1 Methodology
      Bio J. Linnean Society 10 : 199 - 228

      * Geostatistics for Natural Resources by Pierre Goovaerts 1997. Oxford
      University Press.

      Geostatistics and nested ANOVA designs were recommended:

      * You will have to leave the arcview world and enter the geostatistical
      spatial
      statistics world to address these questions. I would start by looking at
      constructing variograms and proceed to kriging surfaces. Variograms model the
      spatial dependence of a feature over space. There are several variogram
      models: linear, spherical, and exponential are the most commonly discussed but
      there are probably others. These models are required to fit the kriging model,
      coefficients from the variograms are used to produce the kriged surface. Apart
      from this you must look at autocorrelation of the two features host trees
      and infected trees. Autocorrelation will tell you the scale over which the
      patterns of infection are correlated.

      * One suggestion was to use a nested design in which samples are taken
      from areas say 10 meters apart and then 100 meters from these samples
      another pair ten meters apart and so forth.

      Point pattern analysis:

      * Point pattern analysis such as mantel tests and K functions were
      recommended. Geostatistics are not appropriate because I am looking at
      points not a continuous pattern and I would be making an assumption of
      similarity. In order to increase the sscale I should map at a larger scale.

      Thank you very much,


      Juliann Aukema


      --
      *To post a message to the list, send it to ai-geostats@....
      *As a general service to list users, please remember to post a summary
      of any useful responses to your questions.
      *To unsubscribe, send email to majordomo@... with no subject and
      "unsubscribe ai-geostats" in the message body.
      DO NOT SEND Subscribe/Unsubscribe requests to the list!