Friday, January 27, 2012

Home is where the Habitat is:

WARNING: This post is full of unanswered questions.  So if you are the type to lose sleep over complex ideas, intertwining webs of confusion, or just simple curiosity then stop reading now.

The 'adventure' that currently consumes most of my life and causes occasional sleep loss is my research project on Mexican spotted owls.  I am currently on my 2nd year of research and gearing up both mentally and physically for a long field season to start in May.  This means I am chugging through historical data entry, preparing sampling methods, selecting sites to visit, and much more.  It is almost too overwhelming to think about everything I have to do.  So instead, I try to focus on the task at hand, otherwise I would surely have a nervous breakdown (which has happened due to uncontrollable circumstances that one hopes they never have to face)!

Why study Mexican spotted owls you ask?  There are plenty of reasons to study this great species.  Firstly, their babies are very cute, as you can see.

To top if off, the adults are a stunning owl with beautiful white spots on sienna colored feathers.  The reason I get paid to do this however, is because these owls are a federally listed threatened species.  Which means, we need to have a well defined and developed recovery plan in order to prevent their status from becoming 'endangered'.

There is one very special aspect about the owls that I study in Utah however.  This is habitat.  Typically Mexican spotted owls (MSOs) utilize forested habitats throughout the rest of their range, including Arizona, New Mexico and parts of eastern Colorado.  Owls nesting in New Mexico and Arizona are typically within stands of pine and use pre-constructed nests that are built by other raptor species (this is because they are smart...why build your own nest when you can use one that is already there?)  The owls in southern Utah however, are not in the forest, but are in canyons.  Furthermore, they do not use stick nests built by other raptors, but instead use caves or cavities within the canyon walls for nesting.  These are called scrape nests.  The female will simply lay eggs on a ledge within the cave or cavity structure and wait for her eggs to hatch.  If you think about it, in both cases the owls are being fairly 'lazy' (or smart in my opinion) by not constructing a stick nest.  What is not clear, is why do the owls in southern Utah inhabit canyons?  How did they get here? Why would they decide to move into these hot, harsh, desert environments when they likely evolved from a dense forested habitat?

Imagine....moving from a nice, shaded forest to a hot, dry canyon.  If you have ever been to Utah, then you may know that it is home to some of the greatest National Parks (or at least that is what I think).  Zion, Canyonlands, and Capitol Reef National Parks are just some of the select parks that this owl calls home.  And of course, since I go where the owls go this is my summer 'office'.

Lost Canyon Overlook, Canyonlands NP photo by Leah R. Lewis
Big Springs, Canyonlands NP       photo by Leah R. Lewis

Capitol Reef NP,     photo by Leah R. Lewis
near Dugout Ranch, photo by Leah R. Lewis
The Needles, Canyonlands NP        photo by Leah R. Lewis

At first glimpse you may not think that you would find spotted owls in this type of environment.  Although hard to find (and trust me you likely would not find them if you tried) they do exist within the deep depths and narrow crevices of these canyonland areas.  My job is to hike to the places these owls reside and find out why they have chosen that select location for nesting habitat.  In other words, why would they be using that particular canyon and not the one next door?  Can I really answer this question just by taking some notes and a couple of measurements? Um...well, I am not really sure.  I will have to let the data speak for itself.  However, one argument that I consistently have with myself is how do we know the owl has selected the 'best' habitat available.  You see, by measuring the habitat surrounding the owls, I am making a very large assumption.  I am assuming that the owl has selected that habitat because it is likely better than the canyon next door.  But is this really true?  Can I really say, hey everyone, owls like these types of narrow canyons with only 10-15 trees and nothing more.  No....not really, because as I know already (since I have been to sites and seen it with my own eyes) these owls are in a wide variety of canyon types.  Meaning they are in wide-long canyons, short-narrow canyons, steep-deep canyons, and some are not in any 'canyon' at all!  Back to my argument, the main question I am continually asking myself is...Are there enough MSOs in southern Utah to occupy all of the best habitat?  My though it NO, definitely not.  Therefore, if I am measuring unoccupied habitat features and comparing those to occupied habitat features is it justified to say that a owl would not like the unoccupied habitat?  Maybe, but maybe not?  Maybe the owl chose the 'occupied' habitat because it was better, but there is also the chance the they chose it because that is what they found first.  Meaning that if they found the 'unoccupied' habitat first maybe that would then be 'occupied' meaning that would then be better then the 'occupied'!  Confused yet?  I know I am.

View from the owls' front door       photo by Leah Lewis
  Inside a nest cave        photo by Leah Lewis

I certainly struggle with whether I should only measure and collect data for habitat the owls are actually using or if I should also look at habitat the owls are 'not' using.  In my opinion, I am not so sure that I should waste my time and energy looking at things owls don't use, but should instead focus on what they do use.   This is where I think my advisor may beg to differ (he says measure everything).  I say...why???

In reality, my project will likely only provide a small insight into the world of these canyon dwelling owls. There are still so many unanswered questions.  I think that many of these questions remain unanswered due to the technical difficulties and supernatural abilities that are sometimes necessary to find and reach these owls.  We can't even say if Utah's population of MSOs are stable, declining, or increasing.  We only have a vague idea of what they are eating within the canyons.  We do not know what they do or if they go anywhere in the winter.  Do they use different canyons in the winter?  Do they move down canyon?  We just don't know.  Why did they come here?  Why did they stay?  Will they evolve into a separate sub-species...or have they already?  Will they continue to persist?

Three juveniles is a 'rare' occurrence.                                                                                            photo by Leah Lewis


  1. Cool stuff, I think you probably should measure habitat features from canyons that they are not using as well. Perhaps you can limit your measurements to nearby areas that are almost certainly within the home range of the owl, which it would have had a chance to assess. This partially circumvents your apprehension about there being too few owls to occupy all the "good" nesting habitat. As we all know, the best recommendation for conservation is not to manage or engineer limited amounts of 'ideal' habitat for a particular species, or even to take a single-species approach at all, but to preserve large tracts of intact ecosystem (yeah right).

    You probably already know, but a method called maximum entropy exists that is very useful in examining habitat preference using only occurrence records (that is, no data on a species' absence from unoccupied areas is necessary). It works by generating principal components from the set of environmental variables you measure at each occurrence point, and examining those that contribute the most to the last principal component (the one that explains the least amount of variation in the data). Theoretically, whatever environmental variable varies the least among all your occupied sites is the most important for site selection by your organism. If they all vary by about the same amount, then there's no evidence for habitat selection, at least with respect to the variables you measured.

    Max ent works well with remotely sensed data. I haven't used it myself, but I know several people who have and I've reviewed some of their papers - it seems pretty powerful.

    Good luck!

  2. Good point Andrew. I will have to keep that in mind. I will certainly be using GIS in my analysis so the maximum entropy could come in handy. I will have to look into that a little deeper. Thanks for the great input!

  3. I agree with Andrew (I think): it is important to measure unoccupied habitat. Without comparing occupied to unoccupied habitat, you have reduced power to tell if owls are choosing certain habitats over others. (I haven't heard of this use of max ent before. It sounds intriguing and intuitively appealing, but I can imagine situations where it could give very misleading results. If given the opportunity, I think comparisons of occupied and unoccupied habitat is a better way to address habitat preferences.) It may very well be that you would find no difference between occupied and unoccupied habitats; that would be support for your hypothesis that owls are not limited by suitable habitat in this area. Or, it may be that there is something that is very critical for habitat selection, but you wouldn't know unless you had measured unoccupied sites.

  4. It's funny, a lot of the original detectability/site occupancy modeling techniques were pioneered for spotted owls in the Pacific NW in the late 80s/early 90s. Since then, theoretical-minded people have taken it in lots of different directions, max ent being but one. There's a big debate over the best process to use for estimating the extent to which burmese pythons in the Everglades can expand their range in North America, using data from their native range in Asia. Max ent predicts that they will not spread beyond south Florida, whereas climate modeling predicts they could one day reach New Jersey and Oregon. You might also want to look into Mahalanobis' distance approaches, these are often used for GIS-based niche modeling of species habitat.

    One good way to know if your model is giving you 'good' results is to let it predict where you'd find owls in an area where you already know the location of the owls, but wasn't included in the creation of the original model.

  5. Well, it is true. I will likely need to measure the 'unoccupied' habitat. I do agree with both of you that a comparison would be good measure. In theory, I will likely use sites that were either never known to be occupied or have not been occupied for at least 3 years. This should allow me to do a paired comparison (or at least I hope). I think that this will give me some indication of 'best' habitat compared to 'mediocre' habitat. I will have to look into the max ent as a side project and see if it can accurately be applied to my situation and within my environment. I will be pleased if I can produce a descriptive outline of potential Mexican spotted owl habitat to biologists across Utah. Surprisingly they do not have anything that definitively describes their canyon habitats.

  6. In addition, you might try to quantify their detectability - that is, how likely is it that, given an owl is present, you will find it in a single visit? Those false absences can be dangerous...

  7. Another graduate student at MSU has already done this as a separate project. He just finished up a couple of months ago and according to his study these owls have a high detectability. In fact it is even higher than northern SPOWs.