Summary of Results

1
Review of Geostatistics in Aquatic Systems Joshua
French and Scott Urquhart Department of
Statistics, Colorado State University Fort
Collins, Colorado
Developments Related to Aquatic Systems
Aquatic Applications
Exploring Spatial Correlation in Rivers
Choice of Distance Metric

Oceans, Seas, and Bays
Author Joshua French Advisor Scott
Urquhart Abstract Semivariograms are used to
explore and quantify spatial correlation of
several particle size and biological variables
for the longitudinal profile of the Ohio
River Data Set The data consisted of between 190
and 235 unique sampling sites (depending on the
variable) collected by the Ohio River Valley
Sanitation Commission (ORSANCO). The data
consisted of both particle size and biological
measurements.
977.5 miles along river
Estuaries
Invalid Covariance Structures Use of in-water
distance leads to invalid covariance
models

• Results
• Exploratory analysis was conducted for each of
  the variables. When reasonable, the method of
  moments empirical semivariogram was calculated
  for each variable. Maximum likelihood was then
  used to model the empirical semivariogram using
  the exponential, Matern, or Gaussian
  semivariogram models. The variogram analysis
  fell into three categories good results, poor
  results, and no results. Good results were when
  the variogram model fit the sample variogram
  reasonably well, poor results were when the
  variogram model fit the empirical variogram
  poorly, and no results were variables for which
  variogram analysis did not seem reasonable.
• Good results
• Percent Gravel, Number of Individuals, Number of
  Species
• Poor results
• Percent Sand, Percent Detritivore, Pecent Simple
  Lithophilic Individuals, Percent Invertivore

Rivers and Streams

Prediction/Estimation Methods

Lakes
Sampling Design and Optimization

Summary of Results
Coastal Systems