#26: Mapping Thermal Heterogeneity & Climate in Riverine Environments

[Dan Isaak]

Pictures really are worth a thousand words, which is why maps are
powerful tools…

Hi Everyone,
I’m not talking about just any old map, mind you, because in this day
& age, given the ubiquity of spatial data & the ease with which lots
of pretty pictures can be made, it’s all too easy to do “GIS tricks” &
call it science. No, I’m talking about smart maps built from lots of
data & the outputs of appropriate analytical procedures that
accurately show the locations and values of the things we care about
and are attempting to manage. Though it’s the 21st Century, I’d argue
that in many regards we’re still trying to catch up to where Lewis &
Clarke were 200 years ago in terms of having maps that show the
current status of stream resources (graphic 1). And if we can’t make
maps that accurately describe status now, it’s going to be really
difficult to describe how things look in the future, let alone
understand what is lost along the way or be able to adapt our
management intelligently to minimize those losses.

So this time out, I thought I’d highlight one of the best tools
available for making maps of thermal conditions in rivers; the
airborne thermal infrared (TIR) system that Torgersen and colleagues
pioneered development of more than a decade ago. The attached paper
has the details, but the system essentially consists of a special type
of camera mounted to a plane or helicopter that is flown along a
section of river. The camera records long-wave infrared radiation
emitted from the water surface, which is calibrated against
temperature measurements taken contemporaneously along the river at
various points. After some post-processing of the remotely sensed
imagery, detailed maps can be generated that show thermal
heterogeneity within the river channel and along the longitudinal
profile (graphic 2).

Although the technology has been around for a while, it’s a
fundamentally important tool for assessing thermal conditions in
larger rivers & streams. Moreover, in streams like some of those we
have in the Northwest where temperatures are already pushing the
limits of what salmon and trout can tolerate, we’re going to need
these inventories to understand where important thermal refugia exist
(& hopefully persist) as temperatures continue to increase. And that
brings us to the second piece of work highlighted in this blog, which
is a recent report commissioned by the EPA to help management agencies
implement temperature water quality standards. The report by
Torgersen, Ebersole, and Keenan is a primer on ways to identify
thermal refugia using a variety of sensor technologies from complex
systems like TIR to simple hand-held temperature probes (graphic 3).
Apparently, this area of research has rapidly expanded rapidly in the
last decade and the report is too large to send as an attachment (18
MB) but it can be downloaded from a webpage here
http://fresc.usgs.gov/research/StudyDetail.asp?Study_ID=540.

More broadly, then, as we continue generating lots of stream and river
temperature data using various sensor technologies, the challenge will
be to integrate & use this data to develop smart maps that show the
current thermal diversity & patterns of climate in streams. The
process of developing those smart maps is often tedious and labor
intensive but these efforts are required to have the necessary
foundation for understanding thermal patterns & accurately translating
various climate change scenarios to stream ecosystems. With accurate
maps, it will be possible to examine differences between historic and
future conditions to see where changes occur and how those changes
relate to important biological entities (e.g., fish population) or
water quality parameters (e.g., a TMDL standard). Where future changes
appear likely to push past important thresholds, we then have to ask
ourselves whether management interventions could make a difference
(e.g., degraded stream) or not (e.g., pristine wilderness stream)? In
the former case, the area becomes a logical candidate for
consideration when choosing where to target limited resources for
habitat and population conservation efforts. The more precisely we can
identify these key areas, the more aquatic biodiversity we’ll be able
to bring with us through this transitional century. There are, of
course, lots of other complicating factors to consider, but at its
core, the essence of the problem is simply picking the smartest things
to do in the smartest places.

So it’s pretty straightforward to see how a tool like TIR could be
used to inventory the most important rivers within a region and derive
accurate maps of current conditions. Perhaps such inventories could
even be repeated in warm years and cold years to describe how
temperatures differ across contrasting climatic conditions to provide
an analogue for future climate change. Yes, that’s all well and good
but it won’t work on smaller streams where the water’s surface is
obscured by riparian vegetation & the costs would be prohibitive if we
tried to do it across the 100,000’s of kilometers that comprise
regional river networks.

To develop smart maps for entire river networks, we need to adopt a
different approach. As we saw in the last blog regarding the NorWeST
database (#25), an informal temperature monitoring network has already
been implemented by hundreds of biologists and hydrologists from
dozens of agencies across the Northwest over the last 20 years. Some
45,000 summers of temperature measurements at more than 15,000 unique
stream sites are represented in that database. Each of those
temperature sites contains some information about the spatial patterns
in stream temperature across the region. All we need is an analytical
technique that allows us to interpolate information between those
sites correctly and then it would be possible to start making accurate
predictions at unsampled locations & to generate smart maps for all
streams at river network and larger scales. The analytical techniques
for doing this sort of interpolation by valid means on networks have
only recently been developed and next time out we’ll start a mini-
module that steps through some of their inherent advantages for
modeling stream temperatures across networks.

Until next time, best regards.

Dan