Understanding Uncertainty in Probabilistic Floodplain Mapping in the Time of Climate Change

Abstract

An integrated framework was employed to develop probabilistic floodplain maps, taking into account hydrologic and hydraulic uncertainties under climate change impacts. To develop the maps, several scenarios representing the individual and compounding effects of the models’ input and parameters uncertainty were defined. Hydrologic model calibration and validation were performed using a Dynamically Dimensioned Search algorithm. A generalized likelihood uncertainty estimation method was used for quantifying uncertainty. To draw on the potential benefits of the proposed methodology, a flash-flood-prone urban watershed in the Greater Toronto Area, Canada, was selected. The developed floodplain maps were updated considering climate change impacts on the input uncertainty with rainfall Intensity–Duration–Frequency (IDF) projections of RCP8.5. The results indicated that the hydrologic model input poses the most uncertainty to floodplain delineation. Incorporating climate change impacts resulted in the expansion of the potential flood area and an increase in water depth. Comparison between stationary and non-stationary IDFs showed that the flood probability is higher when a non-stationary approach is used. The large inevitable uncertainty associated with floodplain mapping and increased future flood risk under climate change imply a great need for enhanced flood modeling techniques and tools. The probabilistic floodplain maps are beneficial for implementing risk management strategies and land-use planning.

The Cost of Imperfect Knowledge: How Epistemic Uncertainties Influence Flood Hazard Assessments

“…Parameters' distributions can be obtained using nominal probability models from expert's knowledge (Kalyanapu et al., 2012; Stephens & Bledsoe, 2020) or statistically calibrated ones generally via the GLUE methodology (G. T. Aronica et al., 2012; Di Baldassarre et al., 2010; Kiczko et al., 2013; Romanowicz & Kiczko, 2016; Zahmatkesh et al., 2021). In this work, epistemic uncertainty will be represented by probability distributions in the roughness parameters only, for the floodplain and for the channel, considering all other parameters, such as DEM or channel bathymetry, as constant regarding the calibration procedure.…”

“…That is, defining a distribution for the uncertain discharge for a given return period. This includes accounting for statistical fitting errors due to limited‐length data and distribution family (Apel et al., 2008; G. T. Aronica et al., 2012; Neal et al., 2013; Romanowicz & Kiczko, 2016; Stephens & Bledsoe, 2020; Winter et al., 2018), additional forcing variables such as flood volume (Candela & Aronica, 2017) or sea‐level rise (Muñoz et al., 2022), or more general hydrograph shape uncertainties through hydrological modeling (Ahmadisharaf et al., 2018; Falter et al., 2015; Grimaldi et al., 2013; Meresa et al., 2021; Zahmatkesh et al., 2021). Others have focused on including uncertainty in the inundation model through its most sensitive parameters such as roughness coefficients (G. T. Aronica et al., 2012; Bharath & Elshorbagy, 2018; Di Baldassarre et al., 2010; Kalyanapu et al., 2012; Kiczko et al., 2013; Romanowicz & Kiczko, 2016), Digital Elevation Maps (DEM) (Apel et al., 2008), or cross‐section geometrical properties (Stephens & Bledsoe, 2020).…”

“…Others have focused on including uncertainty in the inundation model through its most sensitive parameters such as roughness coefficients (G. T. Aronica et al., 2012; Bharath & Elshorbagy, 2018; Di Baldassarre et al., 2010; Kalyanapu et al., 2012; Kiczko et al., 2013; Romanowicz & Kiczko, 2016), Digital Elevation Maps (DEM) (Apel et al., 2008), or cross‐section geometrical properties (Stephens & Bledsoe, 2020). Furthermore, many of these have included both the epistemic uncertainties in the discharges recurrence as well as in the inundation model (Apel et al., 2008; G. T. Aronica et al., 2012; Bharath & Elshorbagy, 2018; Di Baldassarre et al., 2010; Kalyanapu et al., 2012; Kiczko et al., 2013; Romanowicz & Kiczko, 2016; Stephens & Bledsoe, 2020; Zahmatkesh et al., 2021).…”

“…This review of the literature on uncertainty quantification in flood hazard assessments shows that epistemic uncertainties are always kept separate from aleatory uncertainty. As a result, the hazard output is always in the form of a distribution of the probability of exceeding an IM level; whether it is given as a “probability of flood map” for different return periods (Bharath & Elshorbagy, 2018; Di Baldassarre et al., 2010; Domeneghetti et al., 2013; Kiczko et al., 2013; Neal et al., 2013; Stephens & Bledsoe, 2020; Zahmatkesh et al., 2021), or as confidence intervals of IM exceedance curves for each point in space (G. T. Aronica et al., 2012; Nuswantoro et al., 2016; Romanowicz & Kiczko, 2016). While this separation is useful to identify ways of reducing uncertainty, it has been shown to be challenging for proper communication and interpretation, even leading to inconsistent decision making (Der Kiureghian & Ditlevsen, 2009; Fawcett & Green, 2018; Fenton & Neil, 2013; Fox & Ülkümen, 2011).…”

“…The typical outcome from most of these approaches is in the form of "probability of flood" maps for different return periods. That is, for a specific return period, different forcing events and/or inundation model parameters are randomly sampled and used to obtain an ensemble of flood maps from which the probability of flooding is computed empirically (Bharath & Elshorbagy, 2018;Di Baldassarre et al, 2010;Domeneghetti et al, 2013;Kiczko et al, 2013;Neal et al, 2013;Stephens & Bledsoe, 2020;Zahmatkesh et al, 2021). However, a flood risk analysis requires estimating potential damages from the hazard outcomes, and this type of input is not very helpful since most flood damage models use as input the water depth above ground level (and eventually flow velocity and flood duration) (Galasso et al, 2021;Mohanty et al, 2020;Nofal & Van De Lindt, 2022;Pregnolato et al, 2015).…”