Climate change has and is projected to continue to alter historic regimes of temperature,
precipitation, and hydrology. To better understand the combined impacts of climate change
from a land management perspective and spatially identify where the most extreme changes
are anticipated to occur, we worked in collaboration with United States Fish and Wildlife Service managers to develop a
climate change vulnerability map for the Midwestern United States. The map is intended
for use by regional administrators to help them work cross programmatically to prioritize
locations needing support for adaptation planning and for managers to help them
grapple with the impacts projected climate scenarios have on the hydrology of management
units as they develop adaptation strategies. The vulnerability map is watershed-based
(Hydrologic Unit Code-8) and combines fifteen climate change impact and five adaptive
capacity metrics that were selected by United States Fish and Wildlife Service natural
resource managers based upon known and anticipated impacts to species and habitats.
Climate Change Models and Scenarios
We used two Representative Concentration Pathways (RCPs)
, RCP 4.5 & 8.5. RCP 4.5
represents a scenario where greenhouse gas concentrations increase more slowly and decline
around mid-century as a result of climate change mitigation strategies. RCP 8.5 represents a
scenario where no mitigation policies are employed, and greenhouse gas emission continue to
increase through the end of the century. These two RCPs encompass the likely trajectory of end of century radiative
We used five climate models that were Localized
Constructed Analogs (LOCA)
downscaled datasets from phase five of the Climate Model
Intercomparing Project (CMIP5). The models were selected using the EPA’s
Locating and Selecting Scenarios Online (LASSO)
to visualize the variation in the suite of LOCA CMIP5 climate models for US EPA
Regions 5 and 7. The criteria we used in selecting climate models was to capture the range in
projected changes in annual temperature and precipitation across regions and RCPs. The five
LOCA downscaled climate models we selected were:
. Grid cell size was roughly 36 km2, and values were averaged across each watershed.
We used mid-century as our future period (2040-2059), and 1986-2005 for our historic period.
We used the Hydrologic and Water Quality
version 1.1 to run watershed models that simulate hydrologic, sediment,
and nutrient processes using inputs of land-use, land cover, soil type, topography, weather,
and point sources of nutrients. The HAWQS platform is an online tool developed by Texas A&M
University and the United States Environmental Protection Agency to allow decision-makers and
researchers to run large-scale watershed simulation models using the Soil & Water Assessment
Tool (SWAT) model without the need to download/install software, gather input data, perform
initialization steps, or use up local computer resources.
The exposure category had 15 indicators of projected changes in climate with five
indicators in each of three categories: hydrology, precipitation, and temperature (see
table below). Where possible, we used metrics that have been previously defined in the
literature (for example: ETCCDI/CRD Climate Change Indices
). The indicators were selected by
managers and researchers working for the US Fish & Wildlife Service in the Midwest United
States. To capture as broad of a characterization of climate change as possible, we
inspected annual and seasonal changes in temperature, precipitation, and hydrology patterns
across the region in order to select seasonal indicators (e.g. fall mean temperature,
decrease in summer precipitation, increase in spring discharge) that were projected to show
the directional change and captured different periods of the year. If we were unable to
produce a desired indicator or were concerned about the quality of an indicator we
substituted an alternate indicator that captured a similar aspect of climate change or, if
necessary, made a replacement.
The exposure indicator values were calculated as the percent change of each indicator from
the baseline period to future period for each of the five climate models. When calculating
percent change in temperature, the indicators (i.e. AMT and FMT) were converted from Celsius
to Kelvin prior to performing percent change calculations so that the zero point was not
arbitrary. To ensure that all the exposure indicators were positively correlated with
vulnerability we changed the sign (multiplied values by -1) of two indicators: growing
season start (GSS) and decrease in summer precipitation (SUP).
Adaptive Capacity Indicators
Five indicators were identified to incorporate adaptive capacity: density of dams,
landscape diversity, local connectedness, percent cultivated cover, and change in developed
land cover (see table below for links to data sources). We used two layers developed by The
Nature Conservancy (TNC) for their Resilient Lands Mapping Project: landscape diversity and
landscape connectedness. Landscape diversity combines landform variety, coastal lake effect,
and wetland influence into a metric that represents the local variation in microclimates
and microhabitats. Local connectedness is characterized by factors that restrict movement
in the terrestrial landscape (e.g. roads, developed land, agriculture, etc). Density of dams
within watersheds was included as an adaptive capacity indicator to incorporate connectedness
within stream and river channels. To calculate density of dams we used the 2018 National Inventory of Dams from U.S. Army
Corps of Engineers. We used the National Agricultural Statistics Service 2018 Cultivated Layer
raster to calculate the percentage of each watershed that was categorized as cultivated. The
2018 Cultivated Layer uses the last five years (2014-2018) of NASS Cropland Data Layers and
classifies each 30m pixel as cultivated if it was assigned to one of the categories
considered cultivated in at least two years in the five year period or during the most recent
year. Projections of changes in the amount of developed land and urban growth were obtained
from a dataset with land use projections over the 21st century generated using the FORE-SCE
model. We calculated the proportion of each watershed classified as developed during the
historic (average of 1992-1999) and future (average of 2046-2053) period and then calculated
the percent change of each watershed from the historic to future period.
We used the model developed by Glick et al. (2011)
pictured below to conceptualize vulnerability and
used it as a basis for creating a method for calculating a vulnerability index score.
Prior to performing vulnerability calculations each of the exposure indicators and
adaptive capacity indicators were min-max normalized using this formulation: (x-min(x))/
(max(x)-min(x)). The normalized values are then multiplied by the weights selected (left
panel) for each indicator and combined to create the potential impact, adaptive capacity,
and vulnerability scores using the following equation:
is an exposure indicator, Si
is the weight
of that exposure indicator, Aj
is an adaptive capacity indicator and
is the weight of that adaptive capacity indicator.
Vulnerability was calculated for each of the five climate models and then the multi-model
mean and standard deviation was calculated. The composites of exposures (potential impact), adaptive capacity,
and vulnerability were min-max normalized in order to adjust the scale to 0 to 1 to make
them easier to compare.
The underlying data used in this tool is intended for regional comparisons among watersheds.
The hydrology metrics were generated using a regionally calibrated watershed modeling tool.
Additionally, the climate metrics were summarized at the watershed scale. We provide the percent
change from the baseline to the future period values for each exposure indicator for regional
comparisons. We caution against using these values for local planning. A study of a specific
location would require a more robust investigation. Particularly, with watershed models that are
calibrated and validated specific to that location.
The Adaptation Workbook was created by the Northern Institute of Applied Climate Science as
a resource for managers and landowners to assist in the development of management actions
that address climate change. The workbook was initially designed for forest management but
has since been adapted to include other resource types, with more
in the works.
The Federal Adaptation Resources page contains a collection of resources that can aid in
the process of climate change adaptation planning.
Additional Climate Change Projections and Metrics
The National Oceanic and Atmospheric Administration's Climate at a Glance tool allows users to investigate past climate variability and
change at global, national, and more regional and local scales.
The Climate explorer allows the user to create and explore maps and graphs of future and
historic climate changes for counties in the contiguous United States.
Tools and Maps
The Daily Erosion Project map displays the results of an erosion precipitation model that
updates daily based upon the current soil and crop conditions and rainfall amounts at the
HUC-12 scale. Erosion estimates can be visualized by single day or over user selected time
FishTail is a decision support mapper that provides information to assist in support decision
making to conserve stream fish habitats. Users can visualize various metrics on the current
condition of fish habitats as well as future projected conditions at a variety of scales.
The Nature Conservancy’s Floodplain Prioritization Tool allows users to detect areas where
floodplain protection and restoration opportunities exist in the Mississippi River Basin.
The Nature Conservancy’s Resilient Land Mapping Tool provides estimates of
resilience to climate change. We used two components of the climate change resilience
categorizations (Landscape Diversity and Local Connectedness) from The Nature Conservancy’s
Resilient Land Mapping Project as metrics of Adaptive Capacity in our vulnerability assessment tool.
The United States Drought Monitor provides weekly drought condition updates along with monthly
and seasonal outlooks.
The National Weather Service’s Climate Prediction Center produces seasonal outlooks over the
coming year. With general predictions of whether precipitation or temperature will be above
or below normal at different periods of the coming year, these outlooks can be useful when
planning certain management actions.