Detailed Guide to Variables

(for an abbreviated guide - see Quick Reference for Data Entry)

Explanatory variables

These are the variables that we will analyze in terms of their relationship to the richness of invasive species.  For example, do larger reserves tend to have more invasive species?  Do reserves with higher native species richness have more invasives?  These are the kinds of questions we will be attempting to answer.  To do that, we’ll need a good set of explanatory variables to examine.  Following are the main variables we’ll be looking at, in no particular order.

1) Reserve area (in hectares)

All things being equal, one might expect larger reserves to have more invasive species.  Large size could be important for several reasons.  For one, larger reserves are a bigger target.  Many invasive plants have seeds that are wind-dispersed, so having a larger area could make it easier for invasives to establish within the reserve boundaries.  Additionally, if establishment requires habitat disturbance or chance availability of space or resources, a greater area could make for a greater chance that the appropriate conditions would occur somewhere in the reserve.

However, it’s also possible that reserve area per se would not be all that important.  If there are generally a lot of invasive species around (often referred to as “high propagule pressure”), then target size or chance events might not really matter.  In these cases, it would be the specific habitat conditions (see below), not reserve size, which would determine the number of invasive species found.

2) Richness of the “invasive species pool.”  (number of species)

One can think of invasive species in a refuge as being drawn from a larger pool of invasive species that are found within that region.  Some areas may have large numbers of invasives nearby.  If there’s a strong component of chance to invader establishment, reserves in areas with large invasive pools should have more invasive species established.  Viewed another way, if there aren’t many species in the regional pool, then any individual refuge could only have so many invasive species.  For these reasons, any analysis of the number of invasive species within a reserve should try to account for the number of invasive species in the regional pool. 

What constitutes the “pool” for any given refuge is not necessarily clear cut.  Ideally, one would consider the spatial locations of all possible invasive species in the region and whether they are close enough that they could potentially disperse (via seed or humans) to the reserve.  One would also consider habitat affinities – a nearby invasive that lives only in wetlands wouldn’t establish in a reserve that contains no wetlands.  Unfortunately, data with this much detail on habitats and spatial locations are rarely available.  So, we’ll take the invasive species pool to be the exotic plant list for the county (or counties) in which the refuge is located.  This is somewhat course, but it will certainly differentiate regions that have many potential invasives (e.g. Florida) from regions that have few (e.g. Alaska). 

3) Native species richness.

Ecological arguments hold that all things being equal, native species richness should be inversely related to the number of invasives that can establish.  The basic idea here is that there are a limited number of potential niches for plants, so if there are a lot of native species, most of those niches will be filled before invaders arrive.  In contrast, if there are only a few native species present in an area, there may be unfilled niches that invasive species can take advantage of.  The idea that high native species diversity can help prevent the establishment of invasives is often called “biotic resistance.”

The relationship between native richness and invasibility typically holds up in small-scale ecological experiments.  For example, if you remove some native species from a small area, invasives will generally be more likely to establish as compared to control areas in which no native species are removed.  However, in large-scale observational studies in which areas that contain high and low native species richness are compared, one usually finds the opposite relationship – areas with more native species also tend to have more invasive species.   Why would the relationship between native and invasive richness depend on the size of the area being examined?  Many ecologists think the main issue here is habitat quality and heterogeneity.  Note that the argument for an inverse relationship between native and invasive richness is qualified by the statement “all things being equal.”  This may be true in small-scale experiments, but when comparing different sites over large geographic areas, all things are definitely not equal.  In particular, areas with lots of habitat heterogeneity – i.e. lots of different habitats and lots of different niches – will tend to have high native species richness and will also tend to harbor lots of invasives.  So, on a large scale, one will tend to find a positive relationship between native and invasive richness, even if on a small-scale, high native richness can prevent some invasives from establishing. 

Does this sound hopelessly complex?  It’s actually not as bad as it might seem.  In principle, modern statistical approaches can distinguish the direct negative effects of native richness on invasibility (i.e. biotic resistance) from the correlated positive effects of habitat heterogeneity (or area for that matter).  We say “in principle” because whether or not one can actually do this depends on the size and quality of the dataset.  With our study, we will have a lot of data (200-300 reserves), but the quality of these data will be less than ideal.  So, we will make an attempt to separate a direct effect of native richness on invasive richness from the correlations between them (see ANALYSIS), but we won’t really know if we’ve been successful until we have all the data for all the refuges.

4) Habitat heterogeneity (habitat richness and evenness).

Some Wildlife Refuges just protect a single habitat type – a tallgrass prairie, a wetland, a coniferous forest.  Other Refuges contain a variety of habitats, perhaps stretching from a coastal area to an upland forest, with marshes and salt flats in between.  One would expect habitat heterogeneity (the diversity of habitats within an area) to be a strong predictor of the number of invasive species within a refuge.  Really, the appropriate question is not “is habitat heterogeneity important?” (because it almost certainly is), but rather “what aspects of habitat heterogeneity are the best predictors of invasive richness?”

We will examine habitat heterogeneity/diversity in a few different ways.  First, we will divide habitat diversity into two components: 1) habitat richness - the number of different habitats within a refuge, and 2) habitat evenness – the equitability of the distribution of different habitats.  It should be obvious that the number of different habitats is important – a Refuge with eight different habitats will have more invasive species than a Refuge containing only one or two habitat types.  But evenness may also matter.  Consider landcover in the following two made-up refuges, each of which contain the same four habitat types:

Smoky Pines Wildlife Refuge: 500 ha mud flats, 400 ha wetland, 600 ha deciduous forest, 500 ha scrubland.

Great Auk Wildlife Refuge: 5 ha mudflats, 5 ha wetland, 1960 ha deciduous forest, 20 ha scrubland.

Each of these refuges has a habitat richness of n=4 and each is exactly 2000 ha.  But it’s easy to see that the first refuge contains a roughly equal distribution of the four habitats, whereas the second is almost entirely deciduous forest with only small pieces of the other 3 habitats.  One can easily imagine that the first refuge would be more at risk from invasive species than would the second.  For this reason, it is important to think of habitat heterogeneity both in terms of the numbers of habitats and the equitability of those habitats.  Our analysis (see Analysis) will attempt to separate the two.

The second consideration with habitat heterogeneity is that some habitats are more different than others.  For example, wet and dry grasslands may be more similar to one another than either is to an evergreen forest.  Thus, simply counting the number of different habitats may not be an ideal approach to measuring habitat richness. 

Most land management agencies (including the U.S. Fish and Wildlife Service) take this issue into account by using a hierarchical classification of habitat types.  In this scheme, habitats are classified at broad levels (e.g. forest, grassland, open water), and then are classified again within each of these broad levels (e.g. evergreen upland forest, deciduous upland forest).  So, one can actually calculate habitat richness and evenness at both levels of this classification.

The data in the survey for habitat and landcover look promising in that habitats are often carefully divided into many different categories with a specific acreage for each.  Bear in mind, though, that these categories are in many cases arbitrary constructs.  Almost every forest contains a few evergreens – at what point does a forest become a mixed evergreen deciduous rather than just a deciduous forest?  The hierarchies are also debatable.  For example, annual forb habitat and perennial forb habitat are separated at the broadest level of habitat classification, but deciduous forest wetland and deciduous forest desert are within the same category.  Ignoring the fact that you’ve probably never heard of a deciduous forest desert (I certainly haven’t), it’s not all that clear why the first two would be considered more different than the second two. 

You’ll note that a lot of refuge managers themselves express difficulty using the classification.  At Dungeness Wildlife Refgue (http://www.nwrinvasives.com/refuge_data.asp?org=13539) they can’t figure out what do with coastal dunes.  A number of refuge managers don’t know how to classify rocky islands in places like Oregon and Maine.  We won’t attempt to redo all the classifications – that would only lead to additional problems.  Rather, we’ll take the classifications as given, but recognize that they describe only some of the important aspects of habitat heterogeneity.  Ultimately, no one classification scheme would please everyone, and these schemes were designed as general descriptors, certainly not with invasive plants in mind. 

5) Elevational range (meters)

Most species have elevational limits – some are restricted to low-lying areas whereas others are high-elevation specialists.  Even if you have unbroken forest habitat stretching from the ocean to the top of a coastal mountain range (i.e. one habitat type), you’ll likely have different species at the bottom than you’ll have at the top.  Thus, elevational range can be thought of as another aspect of habitat heterogeneity.  The greater the elevational range, the more different habitats are likely to be present, and the greater the propensity for invasive species to establish.   Most refuges provide data on the lowest and highest elevations found, so it is easy to calculate elevational range for any refuge as the difference (in meters) between these values.

6) Regional interactions.(not included in regional analyses)

You will be analyzing the factors that promote invasive establishment within your own region.  After all the regions are finished, we will combine the data from the different regions for a continental-scale analysis.  It may turn out that the factors that are most important in predicting invasibility are always the same, or it may be that these factors vary from region to region.  For example, it’s possible that elevational range mattes a lot in the Pacific region but not so much in the Northeast.  Or, it’s possible that refuge area is an important predictor in the Southeast, but that area is relatively unimportant in Prairie regions. 

These kinds of differences would be called regional interactions, i.e. interactions between factors and the region under consideration.  An interaction is really just a situation where the effect of one factor depends on the level of another factor.  So, an interaction between reserve area and region would reflect a situation where the effect of area on invasive richness depends on the region (i.e. differs among the regions).

Interactions like this are important when thinking about the application of our research.  Knowing about interactions helps us determine whether there are indeed general sets of factors that explain the establishment of invasive plant species or whether the critical factors are region-specific.  If you’re a refuge manager, this effectively tells you whether you need to pay attention to what’s happening in other regions in order to understand your own, or whether you’re best off making predictions just from your own area.

Response Variables

These are the variables that we’ll be analyzing, i.e. our measures of invasive richness.

INVASIVE RICHNESS – This variable is just the number of invasive species found in the refuge, and is the main variable that we’ll examine in our research.

Simple, right?  Except of course it’s not simple.  We have two primary sources of data on invasive richness, and these two sources are really giving us two different quantities.  The first data source is the online Invasive Species Survey and the follow-up e-mails that you send to the Refuge Managers.  These lists are usually telling us about the number of invasive plants that are known to be problems on the refuge – that are, for example, taking over wetlands or preventing forest regeneration.  Some Refuges may list 2 or 3 problem invasives, others might list 20 or 30. 

The second source of information is the full species lists that you will get either from the Refuge websites, or from the Biota of North America site.  These lists will include all exotic (non-native) species known from an area.  This includes problem invasives like the ones in the managers’ lists, but also anything else that’s been found there, including things that may have been planted and not reproduced on their own.  A full list of non-natives might be on the order of 100 or 200 species.  There may be some gray areas, but for the most part, it’s pretty easy to tell which list is which because the sizes of the lists are so different.

Which list is “better”?  As you probably suspect, this is also a complicated question.  From a conservation and management perspective, we might be most interested in just the problem species.  From the perspective of community ecology and general ideas about invisibility, we might want to think about the complete exotic species lists.  Fortunately, we won’t have to choose between these.  For each refuge, we’ll try to find BOTH a list of problem invasives AND a list of all exotics.  We can then analyze the factors that explain both of these, and it will be interesting to see whether these factors are similar to one another. 

One caveat - Invasive richness is a somewhat coarse measure of the intensity of plant invasion at any given refuge.  A refuge might have only one or two main invaders, but these species might cover large areas of the refuge.  For this reason, invasive cover – the proportion of any habitat covered with invasives – is also analyzed in some studies of invasibility.  Unfortunately, we don’t have these data available, so we’ll do the best we can with invasive richness. This shouldn’t be too bad, as in general, studies find that higher numbers of invasives are associated with a higher overall coverage by invasives.