View the Project on GitHub SouthForkResearch/CHaMP-Status-and-Trend-Roll-Ups
Spatially balanced sampling designs aim to spread sampling points uniformly across the spatial domain of interest.
The Generalized Random Tessellation Stratified (GRTS) sample selection algorithm was developed by Stevens and Olsen (2004) and has been used for regional monitoring programs, including several water quality programs funded by the EPA, Bonneville Power Administration, and Washington Department of Ecology, to name a few.
In 2012 the Integrated Status and Effectiveness Monitoring Program provided a free workshop on GRTS-based sampling designs and analysis, which was led by Don Stevens, one of the developers of the GRTS algorithm.
Instructor: Dr. Don Stevens
Red Lion Hotel-Portland Convention Center
1021 NE Grande Ave.
Portland, OR 97232
Several key manuscripts are located in GRTS Workshop 2012/BackgroundArticles
Sections of presentations in GRTS Workshop 2012/BackgroundPresentations supported Day 1 topics. We recommend viewing these background presentations in addition to any ISEMP-specific presentations that are listed in the agenda. All presentations in the Background Presentations folder were generated by Don Stevens.
History and origin of GRTS and overview: *GRTS Workshop 2012/Day 1/GRTS_grts_ISEMP.pptx
Stratified and variable probability sampling
Inclusion probabilities
Site evaluation
Weight adjustments
Introduction to stream network sampling as continuous, discrete or area based samples and the implications of how a response design is set up in these various scenarios
Limitations to application of GRTS along linear networks: GRTS Workshop 2012/Day 1/Sampling a Stream Network.pptx
Specifying objectives: What do you want to estimate? How does design tie to objectives?
Function of the response design and relationship to spatial/temporal design
Spatial design: target population, frames, stratification
Temporal patterns—panel designs
Sampling details: legacy incorporation, densifying the master sample, power analyses GRTS Workshop 2012/Day 1/Merging Legacy Sites with a GRTS Sample.pptx
Extract sample from a master sample
Generating initial weights, probabilities and densities
Documenting the design and generating a design file: what should it contain?
Now that I’ve got my files together, site evaluations, and site metrics, how do I make inferences based on my design, and what do I get (theoretical meaning of results)?
See GRTS Workshop 2012/BackgroundPresentations Olsen_Day 3 1. Spatial Survey Analysis.pptx
Distinguish design based inference from model based inference
Develop adjusted weights (how is this done, what is considered)
Categorical vs continuous estimates
Estimates and indicators: totals (e.g, how many fish do I have?), densities, frequency distributions
Sampling uncertainty and precision of estimates
Analyzing data using post-hoc stratification (how does this work?)
See GRTS Workshop 2012/Day 2/Nahorniak_GRTS Workshop Presentation.pptx and Starcevish_082312.ppt
GRTS function for linear resources:
Adjwgt function
Cont.analysis function
Cat.analysis function
Don’s specific functions (non-spsurvey version of GRTS code)
Use of GRTS for temporal patterns
Other:
Generating a Master Sample (spsurvey)
Development of web based tool for extraction of samples from master samples: what will it contain and what is its status? GRTS Workshop 2012/Day 2/Wep Tool for.pptx