Using the Raven Hydrological Modeling Framework
to Support Climate Adaptation in Local GovernmentManaged Watersheds
INTRODUCTION

PILOT RESULTS

The Raven hydrological modeling framework is a software package
for watershed modeling. It can be used to aid in understanding the
hydrological behaviour of a watershed in light of changes in land
cover, climate, and soil conditions. The framework’s modular design
makes it very flexible, allowing modelers to apply algorithms that
accurately simulate conditions in the watershed in question and/or
test a variety of possibilities for the future.

Mo de l de ve lo p me n t

As a result of the current and future impacts of climate change,
local governments in the Columbia Basin-Boundary region
are facing uncertainty related to the future of the watersheds
surrounding their communities. An accurate and easy-to-use
watershed model could serve as an effective decision support
tool for local government management of drinking water assets
(including natural assets) or land use in watersheds. To test the
value and utility of the Raven framework for local governments in
this region, we undertook a pilot project for the City of Cranbrook.
The City’s drinking watersheds were modeled in order to develop a
preliminary understanding of how projected climate changes and
various wildfire scenarios could impact flow conditions.

The project team successfully built and calibrated a model that
effectively simulated the Gold Creek watershed. It was not possible
to model the Joseph Creek watershed due to questions about the
discharge data available for that creek – the modeling process
identified that either a large volume of water is being lost from the
channel, or that there is an error in how the data are generated.
The project applied two wildfire scenarios (a disbursed burn of 30%
of the watershed and a concentrated burn of 30% of the watershed)
and two climate change scenarios (based on low and high global
emissions).
Mo de li n g re sult s

The outputs of the modeling process indicate that landcover
disturbance from wildfire could have an immediate increase on
the magnitude of peak flows from Gold Creek but that the timing
of flows is unlikely to change substantially. Given changes in
temperature and precipitation as projected with climate change
within approximately the next 60 years, it is possible that there
will be a decrease in the magnitude of floods across the full range

of frequencies; however, this finding should be re-evaluated if
a weather generator that more accurately captures projected
future climate variability is applied to the model. In addition, the
timing of peak flows and return to low flows is projected to occur
substantially earlier in the year.

Pro je ct re qui re me n t s

In order to build and calibrate a watershed model, the following
data are required:
•

At least 20 years of accurate daily discharge data from the
watershed

•

A digital elevation model for the watershed

•

Land use data (e.g., vegetation resource inventory data or
satellite imagery)

•

At least 20 years of daily temperature and precipitation
data from a station with a close climate relationship to the
watershed in question. Where this does not exist, it is possible
to use data from a gridded climate data generator.

Fut ure adjustm ents o r re s e a rc h

To support a more detailed understanding of the potential
hydrological response of Gold Creek to changes in climate and
land cover, further research and model calibration is required. In
particular, the weather generator used to predict climate changes
should be adjusted to incorporate a more accurate projection of
changes in climatic variability.
A copy of the full Cranbrook report is available here.
APPLICABILITY OF RAVEN TO THE LOCAL
GOVERNMENT CONTEXT
De ci si on suppo rt

The Raven framework offers valuable decision support to local
governments who wish to better understand the following in the
context of climate change:
•

how various projected climate scenarios are likely to affect
flow conditions;

•

how various wildfire scenarios are likely to impact flow
conditions; and/or

•

how other types of land use change or forest disturbance
(e.g., logging or interface fire risk reduction) are likely to
impact flow conditions.

The requirement for 20 years of high-quality discharge data
in the watershed somewhat limits the utility of this model to
Basin-Boundary local governments given the lack of long-term
monitoring in many areas. While it is possible to use a proxy
watershed to infer flow conditions, this approach adds an
additional layer of uncertainty to the model.
A watershed model should be built and calibrated by
modeling experts; however, local government staff, if trained,
should be able to manipulate the calibrated model to test the
impact of various climate and land cover scenarios.
Training requires a few hours of staff time assuming existing
background knowledge in GIS or coding.

This information could help inform decisions related to flood
risk management, infrastructure upgrades, land use planning/
approvals, or natural asset management.
It is possible through the Raven framework to simultaneously
model scenarios that test how multiple uncertain factors (e.g.,
climate change and wildfire) may interact to influence future
conditions in the watershed; however, models perform best
when the number of uncertain variables is limited. Introducing
additional layers of uncertainty to the modeling process will
reduce the accuracy of the outputs.

This project has been carried out with financial assistance from the Government of Canada
and the Federation of Canadian Municipalities