#30: Recording & Mapping Earth's stream biodiversity from genetic samples of critters
Dan Isaak
Justifying stream fish electricity rodeos…
Hi Everyone,
If you notice the sequence # of this blog, you’ll see we skipped #29, which will be part 3 and completion to the Spatial Statistical Network Models mini-module. When that blog goes out next month, it will announce the availability of new freeware tools for custom applications of the spatial statistical models. But for now we’re skipping ahead to the first blog in the Climate-Aquatics Biology Module.
That being the case, it’s useful to re-emphasize the fact that many of the things we care about and manage for in streams—water quality attributes, fish abundance, presence of certain species—are often sampled in ways that provide reach-averaged, point values of these attributes scattered across stream networks. As such, most of these attributes might be modeled using the spatial models if the data are organized and maintained in georeferenced databases (Blogs #27 and #28). Moreover, it’s going to be important to work collectively to build these databases and make smart maps of these things because there’s no other way the information will exist otherwise. It’s not going to be possible to remotely sense these attributes any time soon, so it’s going to require people physically mucking around in streams to direct measure these things. Once that’s been done in enough places, it becomes possible to make models linking the things we care about to the physical attributes of streams and networks that can be remotely sensed or otherwise accurately described from existing spatial data layers and then smart maps based on data and model predictions become possible.
That said, it’s often rather enjoyable to muck around in streams and one of the more exciting forms of mucking sometimes involves fish electricity rodeos. We fish biologists often make these rodeos sound more official by calling them, “standardized electrofishing surveys for estimating species composition and abundance” but it’s mainly a ploy to be in the woods and chase partially stunned fish with nets through steep streams whilst jumping over the occasional large rock. As strange as it may seem to the uninitiated, there’s quite a bit of this business that goes on. In a few river basins where we’ve pulled together georeferenced database of both fish surveys and stream temperature measurements, there seem to be 2 – 3 times as many fish survey sites as there are temperature sites. If that rule of thumb holds more generally, the numbers would dwarf the number of temperature sites in the NorWeST database (Blog #25) and there might be somewhere around 50,000 unique stream sites that have been sampled for fish across the Northwest U.S. Moreover, there are 1000s of additional sites sampled each year by biologists from dozens of resource agencies and 10,000s of fish being handled.
At these sites, it’s traditionally been the case that we’re most interested in describing whether or not the species that angers like to catch (e.g., trout, bass, catfish) occur and how abundant they are, but give limited consideration to those less sexy aquatic denizens—the small, nondescript fishes, salamanders, toads, and aquatic insects that live in those same places and often constitute the majority of the overall aquatic community. That’s a shame, because I would guess that 100 years from now, one of the things a 22nd-Century fish biologist will care about most is knowing what biodiversity once occurred at specific stream sites. I suspect they’d very much like to be able to compare a map of what’s there in the year 2112 to what was here now to understand how things changed as we passed through this transitional century (just think if we had a map like that from the year 1900 for comparison!). But even now, with all the fish sampling that’s going on in some places, we’re a long ways from having good status assessments, spatial inventories, and archived databases for the majority of aquatic species in the majority of the Earth’s streams that could serve as useful time-capsules for future biologists. And lacking that basic information, how are we going to design effective strategies for conserving this biodiversity?
One impediment to doing more comprehensive biodiversity surveys is simply the traditional goals we’ve had when sampling fish populations. When the goal is to develop abundance estimates for a few target species at a site, it’s quite possible to do, but also a labor- and time-intensive procedure. Often it involves using block nets to close off a section of stream and then making several electrofishing passes through that section to mark, capture and remove individuals of the target species in ways that meet the statistical assumptions of the abundance estimator. As such, we tend to spend a lot of time at a few sites and on a few species and sometimes ignore large portions of river networks and the majority of species. Biodiversity surveys, in contrast, simply require catching individuals from the various species that occur at a site and sampling enough stream habitat to ensure a minimum level of species detection. For this too, there are standardized sampling protocols to apply, but typically this sort of survey can be done more rapidly than an abundance estimate.
With critters in hand, the standard sets of length, weight, and age measurements may be taken, but a sample that contains a vast amount of biological information about a populations at that site—but is often overlooked—is a piece of tissue. The genes in that tissue are the record of a species’ evolutionary history over the course of the preceding 100s – 1,000,000s of years. That set of biological information is many orders of magnitude larger than the tiny bits of information traditionally obtained regarding the size and number of a few of the critters. Moreover, genetic information can be described very accurately with modern instruments, so once the tissues are analyzed, the results can be digitized to a computer’s memory and archived indefinitely. Twenty, 50, or 100 years from now, a biologist of the future will have an unambiguous description of the aquatic biodiversity that occurred at that site in one of Earth’s streams.
Other impediments to biodiversity/genetic sampling in the past have been the cost and inconvenience of collecting and processing tissue samples. Costs are constantly coming down and for some species that have had single nucleotide polymorphism (SNP) markers developed, it now costs as a few dollars per fish or amphibian to run the analysis. SNPs still need to be developed for many species but as the digital/information revolution marches on it will eventually happen. Between now and then, useful tissue archives could be compiled essentially for free during the course of routine monitoring and fish sampling efforts. It might require a bit more time at a site to take tissue samples from non-target species but that cost is trivial relative to the potential biological information gains and the costs of hiring crews and sending them to and from sampling sites. The last issue is simply one of convenience because preserving/storing tissue samples has traditionally meant having to deal with small alcohol vials that were prone to breakage or leaking. A more efficient protocol was recently developed by LaHood and colleagues (graphic 1; attached paper) that simply involves placing fin tissue on chromatography paper. One hundred samples can easily be stored on an individual piece of paper and archiving the tissues (if there’s no immediate plan for analysis) simply involves stacking them for storage somewhere in the office or lab that’s dry and protected from insects and heat.
It may take a few years, but if the collection of tissue samples became a standard part of stream surveys, significant archives of Earth’s aquatic biodiversity would be accumulated. As sample sizes grew, the data might be used in some instances with the spatial models to literally “map” genetic diversity on top of geomorphic and climatic attributes of stream networks to study and better understand these relationships. But there’s also a large and growing list of applications for genetic data in this rapidly advancing field. The attached paper by Schwartz and colleagues describes a few of these, including: DNA-based abundance monitoring, estimation of demographic rates like survival and recruitment, and detection of pathogens to name a few. Another application is presented in the attached paper by Whiteley and colleagues, for estimating effective population size in headwater trout populations. Quite familiar to many biologists is the use of genetic testing to reveal levels of hybridization, but these same tests permit the real possibility, even in this day and age in a place as densely sampled as the Northwest U.S., of discovering and describing new species as researchers in the Northern Rockies are currently doing (Mike Young, USFS, personal communication). And in the not-too-distant future, it may also be possible to extract DNA from stream water samples to assess and monitor biodiversity (Kevin McKelvey, USFS, personal communication), which would have the unfortunate consequence of rendering some fish rodeos obsolete.
Until next time, best regards.
Dan