#38: Part 2, Monitoring to detect climate effects on fish: Resurveys of historical stream transects

37 views

Skip to first unread message

Dan Isaak

unread,

Jan 28, 2013, 1:31:01 AM1/28/13

to climateaq...@googlegroups.com

A “work-around” solution to time travel…

Hi Everyone,

So if only we could, like the protagonist in H.G. Wells classic 1895 book, The Time Machine, build a machine to travel back several decades and survey fish populations along stream transects to establish baselines for contemporary climate change assessments, then we’d be in business. But as of yet, time-travel has proven to be insoluble except in Hollywood & perhaps Einstein-ian physics, so we need a less expensive “work-around”. (As an aside, if you haven’t seen the 1960 movie adaptation of Well’s book, I highly recommend it http://www.imdb.com/title/tt0054387/. Special effects are corny by today’s standards but they were pretty fancy for the times and the story’s a timeless one, but I digress…). A much easier, yet effective solution to time travel is simply resampling species occurrence along transects that were first sampled at least several decades ago. Based on the power calculations for stream isotherm shifts from last time (blog #37), that should ensure enough time has elapsed that distribution shifts related to climate change could occur. The attached study by Moritz & colleagues (graphic 1), where Grinnell’s Yosemite transect from 100 years ago was resampled to document shifts in mammal distributions, is an excellent example. Although it’s not a stream study, the premise is the same. Survey species occurrence along an elevational transect, use statistical techniques to estimate the boundaries of species distributions, and compare the locations of those boundaries between surveys.

Simple enough, and streams, by their linear nature, have lent themselves to transect surveys, so dozens, if not hundreds, of these surveys have been done previously (graphic 2). These streams are now prime candidates for resurvey efforts and I suspect contain the evidence needed to link climate warming to fish distribution shifts. That said, there are potentially confounding factors that will often interfere or synergize with climate trends. A good example is provided by a recent study by Hitt and Roberts (attached) wherein the authors resurveyed Burton & Odum’s Virginia stream transects from ~69 years ago (graphic 3). At many sites, the authors found significant alterations of historical fish communities associated with species invasions and habitat alterations, with communities in downstream reaches most affected. Although Hitt & Roberts is not a climate change study per se, the warming effect is probably there mixed up with everything else given that air temperature across the state have been warming rapidly (http://www.climatecentral.org/news/the-heat-is-on/).

It will not be uncommon for confounding factors to occur, so isolating the effects of warming will take careful study but there are ways. One option is simply to avoid streams with confounding factors and focus on resurveys of native fishes in “pristine” streams not being invaded by non-natives. That’s often easier said than done in many areas these days, but there are instances, especially in mountain streams where the upstream extent of species distributions are delimited by cold temperatures, habitats are in excellent condition, and invasive species absent, where it is possible. Another option is to amp up the statistical power to detect trends by increasing sample size, which would allow the relative effects of habitat degradation, invasive species, and temperature increases to be estimated. So rather than resurveying 1 or 2 streams, we resurvey dozens or perhaps a few hundred where the best historical baselines exist (graphic 4). Remember, we’re already doing thousands of stream electrofishing rodeos each year (blog #30), so targeting even a tiny proportion of these rodeos to resurveys wouldn’t require much in the way of additional resources—only a bit of time to compile the historical site information into a digital database and some time to coordinate resurvey efforts within and among agencies. Dirt cheap, given the wealth of new information we’d learn from developing and formalizing such networks of “sentinel streams.” This concept isn’t exactly new, having been around for many years in the arena of long-term ecological research but the time is now ripe in the fish world given what we need to learn in a hurry. The attached paper by Craine & colleagues calls these coordinated networks EDENs for “Environmentally Distributed Ecological Networks” and covers some of their general in’s and out’s.

So that’s it for now. Next time out, we’ll examine a few additional ways of deriving estimates of distribution shifts from commonly available datasets. These will involve something called “BIG DATA”, which is intimately linked to “Crowd-Sourcing”—a couple of terms and powerful techniques we’ll need familiarity with as we work to unlock the time capsules that many historical datasets ultimately represent.