Project Introduction 2014

Caveats, warnings, and disclaimers

This is real research – it is not a canned exercise for trying to teach you something. We don’t know what we’ll find out, we don’t even know if we’ll find out anything at all. That’s how research goes, for better or for worse.  Before you read anything else, read what follows:

• Problems will almost certainly come up as we get into the data compilation end of the project. We’ve tried to anticipate as many of these as possible, but there are always things that one doesn’t find out until one starts. Be ready to be flexible - this is also part of doing research.
• If you don’t know, or if you’re not sure, just ask. We know you don’t have experience with any of this. If you’re wondering about something it probably means we didn’t explain things very well. Ask your instructor or post a question on the Discussion Board.  But please do ask. I’ll always respond promptly to posted questions on the Discussion Board.   
• Don’t guess! That is, don’t enter data unless you have real confidence in those data. In any scientific research, it’s always better to have missing data than data that are wrong.
• Don’t assume that all the data you find are necessarily correct.  There could be errors in entries within the NAAMP data, or even within the GIS data.  Did 40,000 cars really pass an observer during a one-minute listen for frogs?  Is a road really 2 miles wide?  If you see things that strike you as odd, speak up about them and don’t assume they’re correct. 
• Take notes about questions or issues that come up while you’re doing the data compilation. It's really hard to remember questions from one route to the next. It never hurts to write things down, either on paper or in the “notes” section of your spreadsheet.
• Materials on this website probably contain typos or other mistakes. Furthermore, what you read may contradict something your instructor has told you. If the contradiction is theoretical or conceptual (e.g. what is “habitat fragmentation” or why are some areas more species rich than others), go with what your instructor told you. If the contradiction concerns research protocols (e.g. what data to compile, how to calculate something) be sure to post it on the Discussion Board so we can clear it up.
• Don’t be afraid to voice concerns, objections, skepticism. We’ve tried to think through the limitations of the project and how we’ll deal with them. But you may think of some issues that we haven’t considered. If you do think of some additional problems, we want to know. Being skeptical of one’s own results is an important part of science.
• If you have some ideas for additional analyses you’d like for us to do, additional questions you think we should be asking, please let us know. This project is not set in stone –you may have good ideas that can be incorporated into the project.
• Take the time you need to do things well. You are contributing to an important research project and your contribution matters. Good luck!

About this project

The National Center for Ecological Analysis and Synthesis was founded in 1995 with a mission to promote the sharing and use of ecological data to solve environmental problems. The center promotes science in a number ways – through scientific working groups in which scientists from around the world meet to work on specific problems, through resident scientists, most of whom are fresh out of PhD programs, and through distinguished visiting scientists who come to the center for varying lengths of time.

One very successful NCEAS program has been the “Distributed Graduate Seminar” (DGS) series. In a DGS, a large research project is broken up into pieces and distributed to small graduate courses all over the country. These graduate courses work only on this specific project and then contribute their piece to the overall analysis. Past distributed graduate seminars have examined issues like the scientific basis for endangered species recovery plans, the role of marine reserves in conserving coastal ecosystems, and the economic impacts of forest pests amd pathogens. Many of these projects have led to important scientific findings and had an impact on conservation policy.

The goal of this project is to do something similar at the undergraduate level. We chose this specific project (landscape ecology of amphibian populations) because most undergraduate ecology and conservation biology classes cover habitat loss and fragmentation and because a dataset (the North American Amphibian Monitoring Program) was available and ready to be analyzed. In addition, the methods involved in the study (e.g. Geographic Information Systems) are increasingly accessible for people with little or no prior exposure. 

Your class was chosen because of your instructor’s interest in the project and because of his or her expertise with these issues. We hope you’ll enjoy working on this, and we hope you’ll learn something from the experience. We also hope to get some real science done.

Scientific Objectives

Habitat fragmentation is the division of large continuous habitats into smaller, more isolated units. In many parts of the world, habitat loss and the resulting fragmentation may be the single greatest threat to amphibian populations. Habitat fragmentation may be particularly bad for amphibians because: 1) many amphibians have to migrate between aquatic breeding habitats and terrestrial habitats where they spend the rest of the year, 2) even narrow bands of unsuitable habitat (e.g. roads) can act as barriers to amphibian movement, 3) amphibians are slow-moving and are thus extremely prone to being killed on roads, and 4) unlike many birds and mammals, amphibians usually cannot travel long distances in order to leave areas that have become unsuitable.  The overall goal of our study is to examine the relationship between amphibian distributions (i.e. where amphibians do and don’t live) and characteristics of the surrounding landscape such as forest cover and the presence of roads. 

We began this project in 2013.  With 9 classes working together, we matched frog presence/absence data with land use variables for over 1600 sites across 15 states.  We found that both vehicle traffic and road length were negatively associated with the species richness (i.e. the number of species found) at any given site.  Wetland area was positively associated with species richness (not surprising) but the effects of forest cover and agriculture were highly variable from one species to the next.  Interestingly, once we had accounted for the effects of roads, low to moderate levels of development tended to have a positive association with amphibian richness, probably because humans inadvertently create new breeding sites for frogs (e.g. ditches, garden ponds). 

To keep things simple for the first year of the project, we assessed all these variables (e.g road length, forest cover, development) within 1 km of each survey location.  One kilometer is reasonable - this is about as far as most individual frogs would move - but we may well have missed important effects by choosing an arbitrary scale like this.  It's possible, for example, that forest cover would mostly matter at a much smaller scale (frogs may need forest near their breeding pond but not really be helped by forest 500 m away).  Other factors (e.g. development, wetland area) might be most important at large scales.  For instance, sites with little development within 1 km might not be able to support amphibian populations over the long term if they are isolated by a lot of development at much larger scales.  Road effects could go either way - if the negative effects of roads are largely due to habitat disturbance, roadkill, or the application of deicing salts, then road effects may be observed primarily at smaller scales.  But if the main effects of roads are to isolate populations from one another, these effects would tend to show up at larger scales (e.g. a few kilometers).  So, to really understand how land use affects amphibians, we really need to look across multiple scales. 

In this year's project, the goal is to take factors that were important in last year's analysis (traffic, road length, development, forest cover, agricultural cover, and wetland area) and examine their importance for amphibian presence and species richness at a range of scales from <300 m from each site out to 10 km.  For each of these factors, we will try to determine the scale (i.e. range of distances) over which they maximally influence amphibians. 

We will also introduce a couple of new measures for the 'connectivity' of each landscape - you can think about connectivity as sort of the opposite of fragmentation.  Connectivity measurements assess how easy it should be for frogs to get from one wetland to another or from wetlands to other habitats that they need.  We will compare a connectivity measurement that only considers wetlands to one that considers both wetlands and terrestrial habitats.  This is a way of asking whether one needs to consider the locations of terrestrial habitats when evaluating if a particular area should be suitable for the long-term survival of amphibian populations. 

Project structure

Each of the participating classes will have a set of assigned routes on which amphibians have been surveyed as part of the North American Amphibian Monitoring Program (NAAMP). Usually, these routes will either be within your own state or within a nearby state. These routes will be divided up among the students in the class by your instructor. 

Each route consists of 10 stops where a volunteer has gone out to listen for frog calls.  For each route, you’ll only be expected to compile data for one of these stops.  Because we'll be looking at the landscape within 10 km of each stop, different stops within a route would mostly consist of overlapping landscapes, so they wouldn't really give us independent data.   See: http://www.pwrc.usgs.gov/naamp/index.cfm?fuseaction=app.protocol for a complete description of the NAAMP survey protocols (this is listed under Important Links on the project website).

For each survey stop, there are two kinds of data that you will compile.  The first type is a summary of the frog and toad call data from that stop.  This mostly just requires some manipulations in Excel.  The second type of data is the landscape data from the stop.  This is a little more challenging, but still straightforward.  To compile the landscape data, you will import several different types of landscape data into Geographic Information System (GIS) software and carry out some calculations using these data.  Complete protocols (i.e. instructions) for compiling all these variables are given in the Research Protocols section of the website.  But the Project Outline below gives you a good overview of the steps involved.

Project Outline. 

1. You will be assigned several survey ROUTES by your instructor.  For each ROUTE, we’ll be compiling data from one randomly chosen STOP where people have gone out to listen to frog calls in order to determine which species are present. 

2.  Given any assigned route, job number one is to summarize the frog call data from the route in order to determine how many species (and which species) are present at each stop.  See the protocol: “summarizing frog and toad call data from each stop” for details on how to do this. 

3.  Then, Job number two is to put together information on land use, wetlands, and roads within several different distances from each stop.  This will be done using Geographic Information Systems software (qGIS or arcGIS).  You will create a map using this software and then use the map that you create to make several calculations, including things like the total length of road surface and the total forest area within a given distance from each stop. 

4.  Each route has been independently assigned to at least two students in order to validate the data.  When you have completed steps 2 and 3 for a route, you will meet with the other student who had the same assigned route and compare your results.  Where you came up with different results, you can work with each other, your instructor, and/or me to try to resolve the discrepancy.  When everyone agrees on the results for a route, they can be entered into a database for your class.

5.  When your class has finished all its assigned routes (usually, a few weeks), your class will work together to analyze the data and determine how the landscape variables you measured in steps 3 and 4 affect the amphibian variables that you summarized in step 2.

6.  Representatives from each participating class will travel to NCEAS in April to present the results for their state(s).  We will then work to compile and analyze the data from all the classes put together.  Those who come to the meeting can then present these results to the rest of your class.