National Park Service

Inventory & Monitoring (I&M)

Advanced Topics

Please direct questions and comments about these pages, and the R-project in general, to Dr. Tom Philippi.

These pages present advanced topics presented via webinar between December 13, 2011 and March 27 2012.

Goals

Between Paul Geissler and myself, we have offered at least 6 Webinar series on the basics of using R for natural resource questions. While they have been offered at various levels and speeds, they have all covered the basics of R, but not more advanced topics specific to NPS I&M and Natural Resources in general. One of the major advantages of R is that experts in many fields find R useful for their work, and then make their tools available as packages for climate data, water quality, mark-recapture, distance sampling, or occupancy estimation, and even analysis of sounds. While I have promised more advanced topics following previous courses, by the time the basics series was completed, both the audience and I had run out of energy, and we all had other work to get done, so the advanced topics never quite happened. This winter we will cover a few of the topics and packages more broadly useful for natural resources and especially monitoring.


Assumed Background and Prerequisites

As noted above, these are NOT another introduction to R. In order to get through the material, a basic familiarity with R will be assumed. We will not spend time on reading and writing data via read.csv, simple reshaping, merging, and subsetting of data frames, or simple graphics. Any of Paul or my webinar series would suffice, or even a few hours working through one of the good introductions available at the CRAN website, or the content of Paul's web pages.

Given that background, each of these topics is independent of the other topics. I expect that most folks will choose to participate in just the topics of interest to them. Therefore, none of the topics will assume or require material from other topics, although Thursday sessions will build upon the material in that Tuesday session.


Topics Covered

Date Topic (URL to corresponding webpage) Webinar Recording
Dec 13 Testing for Trends in Site-Revisit Monitoring recording
Jan 10 Spatial Data in R recording
Jan 12 Spatial Sampling (GRTS) recording
Jan 24 Automated Reporting (tables & graphs) recording
Feb 7 Advanced Graphics (lattice) -
Mar 6 Species Composition (manipulations, diversity/richness, classification) -
Mar 8 Species Composition (ordination, hypothesis tests) -
Mar 27 Instrumentation Time Series -

⇑ To Top of Page

Topic Descriptions

Testing for Trends in Site-Revisit Monitoring

One powerful approach to environmental monitoring for status and trends is to draw a probability sample of locations or sites, revisit those sites on a regular schedule, then test the resulting data for temporal trends. Complications can include both random and fixed effects among sites, and response variables that are binary (presence/absence or % cover) or Poisson (counts of individuals) instead of Gaussian. This topic will use glm, lme, and glmer to fit models appropriate for such monitoring data. The most complex designs require RwinBugs or other MCMC approaches to fit; those toolds will not be covered in this topic.


Spatial Data in R

We will import and use both vector and raster data in R, primarily with sp and rgdal.


Spatial Sampling (GRTS)

We will generate different GRTS draws from point, line, and area targets, with and without stratification and panelling, and then extract attributes such as


Automated Reporting (tables & graphs)

We will generate sets of tables and figures (e.g., one for each species) for inclusion in html and MSword documents. Automated report generation and formatting is much easier to do with Sweave, but often one is forced to use other "industry standard" tools.


Advanced Graphics (lattice)

The power of lattice is in presenting clear, concise graphs of complex data. In addition to various presentations (scatterplots, box and whisker plots, wire-frame or contour plots), additional grouping or covariates can be included as colors/symbols within panels, and by multiple panels on a single page, scaled and ordered to emphasize the important comparisons. However, it can be difficult to figure the most effective presentation, and then out how to tweak the different elements in lattice in order to generate that figure. We will work through several examples.


Species Composition (manipulations, diversity/richness, classification)

Tools to reshape species by site by time datasets, linked tables of species by traits, species presences by sites, and environmental attributes by site, and then compute various diversity measures, similarity/dissimilarity, and classifications of sites based on species.


Species Composition (ordination, hypothesis tests)

Due to my personal biases, nMDS and not DCA/CCA ordination will be presented for visualizing patterns in species composition. ANOSIM and adonis will be used to test for categorical and continuous predictors of compositional dissimilarity.


Instrumentation Time Series

Instumentation such as temperature or pH loggers can produce time series of autocorrelated measurements. Rather than computing means, much more ecologically-relevant parameters can be extracted from such data, such as minimum O2, durations above or below a threshold, and changes in the peak and duration of "events".

⇑ To Top of Page

Last Updated: December 30, 2016 Contact Webmaster