The Groundwater Resources Association of California will conduct a one-day seminar on Statistics for Groundwater Investigations on May 7, 2002 (Tuesday) at the Marriott Hotel in Walnut Creek, CA. Dennis Helsel, Ph.D., Geologist with the U.S. Geological Survey, will be the seminar leader. Dr. Helsel received his Ph.D. in Environmental Science and Engineering from Virginia Tech and is co-author of the textbook, Statistical Methods in Water Resources (1992). He has designed and taught training courses on environmental statistics since 1986.

Tips for touring environmental data are remarkably similar whether traveling in one dimension (describing data), two dimensions (plots and regression models), or cruising along 3-D surfaces like kriging. Some roads are smooth, others bumpy. This one-day guided tour stops at some of the best-loved locations, as well as important out of the way spots, to understand the landscape of interpreting environmental data.

Course Outline

Know Your Destination

• Questions of frequency versus total amount.
  
• Identifying which type of question you are asking
  
• Travel guides - planning the trip in advance.
  

Know the Terrain

• Characteristics of groundwater data
  
• Skewness, values below detection limits, others
  
• Data characteristics that are roadblocks to getting there
  
• Graphs: maps for choosing which route to take

Know Your Vehicle

• Hypothesis tests: testing frequencies or amount?
• Characteristics of parametric and nonparametric tests
• Effects of skewness and outliers on hypothesis tests
• Effects of skewness and outliers on confidence and other intervals

Touring Data in One Dimension

• Summary statistics
• Effects of skewness and outliers
• Estimating a mean - one of the most difficult tasks for GW data!
• Transformations - taking an alternate route
• Summary statistics for data with nondetects
• Methods for comparing data to a standard

Touring Data in Two Dimensions

• Methods for fitting lines to data
• Effects of skewness and outliers
• Effects of transformations
• Regression lines versus smooths
• Fitting lines with nondetects

Touring Data in Three Dimensions

• Once again, frequencies or amounts?
• Effects of skewness and outliers
• Methods for fitting surfaces to data
• Effects of transformations
• Kriging versus smooths