QUEST
Tutorials & Workshops
Tutorials
Spatial Interpolation (Kriging)
- With ArcGIS, tips from Harvard Univ.
- A good background!
- Help from ArcGIS
- Spatial Interpretation handout, used with permission, Geography Department, Colgate University
Monte Carlo and bootstrapping
Trend testing
Thank you to Ruth Yanai's spring 2015 seminar class for reviewing these for us!
Uncertainty Workshops
Module 1: Measurement uncertaintyPresenter: Hank W. Loescher and Janae L. CsavinaThis module will provide an introduction to the GUM (Guide to Uncertainty Measurement), which provides a common basis for quantifying uncertainty in measurements of any type. The GUM was developed by the JCGM, the Joint Committee for Guides and Metrology of seven international organizations (BIPM, CEI, IFCC, ILAC, ISO, UICPA, UIPPA et OIML). Topics will include: use of international and nationally recognized standards, steady state calibrations, field validations, managing uncertainty, automated quality control activities, and sensor measurement schemes. Participants will work with examples of uncertainty assessments applied to various measurements. |
|
Module 2: Experimental Design for Long-Term MonitoringPresenter: Christina L. StaudhammerThis module will focus on designing appropriate and effective observational and manipulative studies, including both plot-based sampling and high-frequency sensor applications. We will examine the basic requirements for rigorous statistical testing (randomization, replication, independence), and investigate how design control can be effective in improving efficiency of data collection efforts. Quantifying uncertainty provides a basis for optimizing designs for environmental monitoring designs to best make use of limited resources.
|
|
Module 3: Monte Carlo Error PropagationPresenter: Oswaldo Carrillo and Ruth D. YanaiThis module will show how to use a Monte Carlo approach to estimating uncertainty, using Excel and R. Examples include estimates of forest biomass and nutrient content, which require propagating error in tree measurements, regression models, and mean concentrations. Participants should bring laptop computers and ecological data and calculations in need of uncertainty analysis (you can use ours if you don't have your own). At the end of the workshop, some participants will have documented the uncertainty in their result. All participants will understand the principles of Monte Carlo sampling and will have tools for implementing uncertainty analyses. |
DOWNLOAD R |
Module 4: Uncertainty quantification: analysis of NEON and other biodiversity network data with hierarchical BayesPresenter: James S. ClarkBayesian methods provide a natural framework for quantifying uncertainty in data, parameters, models, and predictions. This session will summarize basic concepts for hierarchical modeling. Lecture materials and code will be available in advance, with most of the session devoted to hands-on implementation of basic models and implementation. Participants are encouraged to bring their own data sets. |
DOWNLOAD R |