#20: GIS tools for mapping streamflow responses to climate change across the western U.S.

[Dan Isaak]

Planning for the future requires seeing the possibilities now…

Hi everyone and happy new year,
So as was alluded to in the last blog, this time out we’re
highlighting a few of the tools and GIS outputs that are available for
visualizing the sorts of hydrologic responses to climate change that
we’ve been discussing through the series of blog posts in the climate-
aquatics hydrology module. In a sense, these tools are like the air
temperature/climate change mapping tools that were the subject of Blog
#13, but take things out of the atmosphere and into the stream by
translating air temperature trends to effects on streamflows by
running different climate conditions through a hydrologic model. These
sorts of tools are a great way of making the science readily available
and can serve as powerful planning & coordination tool by bringing
various stakeholders together within a common framework. We’ll no
doubt see more of these tools in future years given the ease with
which they are now developed and because new models are increasingly
coming online that are linked to spatial datasets and real-world
coordinate systems.

Both sets of tools we’re looking at today are derivatives of outputs
from the Variable Infiltration Capacity (VIC) model, which was first
developed in the early 1990’s and has since been used extensively to
model stream hydrology at broad scales across many parts of North
America. VIC is a mechanistic land-surface hydrological model that
simulates runoff generation within a specified area like a grid cell
in a climate model or digital elevation model as a function of the
relevant physical processes. For more details on those processes or
the guts of the model, please consult your friendly local hydrologist
as most are probably familiar with it or consult some of the key
references associated with VIC model development and application
(graphic 1).

The first tool that maps VIC outputs was developed by the Climate
Impacts Group and a consortium of partners across the Pacific
Northwest and provides means of accessing and downloading historic and
future hydrologic scenarios at 300 locations on large rivers (graphic
2; http://www.hydro.washington.edu/2860/). Future climate scenarios
were developed using outputs from 20 global climate models run with
the B1 (moderate warming) and A1B (faster warming) greenhouse gas
emissions scenarios developed by the Intergovernmental Panel on
Climate Change (IPCC). To access the data for a specific river site,
simply select a site through the convenient GoogleMap interface and
you’ll be taken to a webpage hosting maps, graphs, and data summarized
at various time-steps for download and subsequent use.

Pretty darn slick to be able to see future climate scenarios on a
local river at the click of a button, but the large rivers that the
preceding tool was developed for ultimately comprise only 5% - 10% of
the stream length within a regional network. What about all those
other rivers and streams, which resource agencies also need
information on? That problem was tackled by Wenger and colleagues
(attached pdf), who translated the VIC model results to individual
reaches in the 1:100,000-scale NHD+ hydrologic layer that the USGS has
developed for the entire U.S. (http://www.horizon-systems.com/nhdplus/
data.php).

Because it was rare to have used VIC like this previously, the authors
went to some lengths to validate the model predictions against
observed flow data at 55 stream gage sites in the Pacific Northwest
(graphic 3) before the model was used to forecast climate-related
hydrologic changes. The validation work suggests VIC predicts things
like the timing of runoff and the frequency of high winter flows
relatively well but doesn’t do as well with other metrics related to
summer flows for example. The validation work suggests VIC predicts
things like the total annual flow and the frequency of high winter
flows consistently well over the region but doesn’t consistently
perform as well with other metrics related to extreme high or low
summer flows for example. Annual runoff timing showed an interesting
and correctable bias with rain dominated locations having late
hydrographs while snow dominated locations had early hydrographs,
sometimes on the order of a month. The validation work also plays an
important part in improving the VIC model, and future versions will
benefit from the insights gained by looking at performance across
these 55 stations.

Once the VIC model’s performance on smaller rivers and streams was
understood, it was then used to describe historical and future
hydrologic patterns across much of the western U.S. (graphic 4) using
the same A1B scenario that fed into the Climate Impacts Group scenario
tool above. The outputs of these model runs are stored as table
attributes that are easily linked to the NHD+ hydrolayer and it’s all
available on the “Western US Stream Flow Metric Dataset” website
(http://www.fs.fed.us/rm/boise/AWAE/projects/
modeled_stream_flow_metrics.shtml); (a quick-start guide on
downloading the NHD+ hydrolayer prior to adding the flow metrics data
can be found here:
http://www.fs.fed.us/rm/boise/AWAE/projects/stream_temp/multregression/downloads/2_N_HowToDownloadStreamsUsingNHDPlus.pdf)

As with any model, there’s a certain amount of “buyer beware” and it’s
necessary to kick the tires and fully understand what the strengths/
weaknesses are before using it as a basis for decision making, so
please read the metadata on the website. Even with the warts, however,
the Western Flow Metrics Dataset derived from VIC serves a useful
function by providing a nearly continuous spatial map of several
important hydrologic characteristics. That makes it possible to
understand how climate change may affect hydrology throughout entire
river networks and to understand the context associated with
individual points in those networks, which are often necessary for
planning efficient conservation strategies. These continuous maps also
make it possible to do apples-to-apples comparisons across broad areas
or to clip the VIC flow metrics to any jurisdictional boundary (e.g.,
a national forest, a BLM district, a fish & game management unit,
etc.) and develop custom climate assessments for those areas.

So lots of useful models, tools, and descriptive studies out there for
understanding & predicting how climate change is/will affect stream
hydrologies (blogs 16 – 19). We’ll wrap up the climate-hydrology
module next time out with a few concluding thoughts, then will move
into the biology module where the fun will really start. Before making
the shift, however, we’ll take a short break from our regularly
scheduled programming to do one or two ‘house-keeping’ blogs that will
keep us current on recent research regarding stream temperature
responses to climate change and we’ll also do a ‘mini-module’ to
present some powerful new analytical techniques and programs for
spatial statistical analyses of data on stream networks that are
becoming available and can yield fundamentally more accurate
information for climate change and other stream assessments.

Until next time, best regards,
Dan