Lidsay Flood 2014 Daily Close Read Practice Answers

Abstract

Floods are the nigh mutual and widely distributed natural hazard, threatening life and belongings worldwide. Governments worldwide are facing significant challenges associated with flood hazard, specifically: increasing urbanization; against the background of doubt associated with increasing climate variability nether climate change. Thus, inundation hazard assessments demand to consider climate change uncertainties explicitly. This paper explores the role of climate modify doubtfulness through dubiety analysis in flood modelling through a probabilistic framework using a Monte Carlo approach and is demonstrated for case report catchment. Different input, structure and parameter uncertainties were investigated to empathize how important the office of a not-stationary climate may exist on future extreme overflowing events. Results suggest that inflow uncertainties are the almost influential in order to capture the range of doubtfulness in alluvion extent, more important than hydraulic model parameter uncertainty, and thus, the influence of not-stationarity of climate on alluvion extent is critical to capture. Topographic controls are shown to create tipping points in the inundation–flow relationship, and these may be useful and important to quantify for future planning and policy. Full Monte Carlo analysis within the probabilistic framework is computationally expensive, and there is a need to explore more fourth dimension-efficient strategies which may outcome in a like estimate of the total uncertainty. Unproblematic uncertainty quantification techniques such as Latin hypercube sampling approaches were tested to reduce computational burden.

Introduction

Floods are the well-nigh common and widely distributed natural hazard, threatening life and property worldwide (Jonkman and Vrijling 2008). Flood take chances is a function of flood hazard and issue (IPCC 2014; Balica et al. 2013). Flood hazards consequence from many unlike sources (e.thou. coastal, fluvial, pluvial or estuarine), whilst the consequences arise from the adverse impacts of flooding on people, property, human health, the environment, cultural heritage and economic activity (Beevers et al. 2016). The Un estimates that 1 Bn people live in areas of potential flood risk and harm caused by floods on a global scale has been meaning in contempo decades (Jonkman and Vrijling 2008). In the last decade (2007–2017), there take been around 200 significant alluvion events in Europe, resulting in almost 1000 deaths, affecting 3.9 M people and causing over $55 Bn worth of damage (2018). In the UK, the estimated cost of flood damage in the UK was £3.2 M, whilst the 2013/2014 events are estimated to have cost the economic system £1.three K. The UK government estimates that each year flooding will cause £ane.i Bn, whilst maintaining current levels of flood defence will require a farther £ane Bn per year past 2035 (Sayers et al. 2015). Looking forward, governments worldwide are facing significant challenges associated with overflowing run a risk, specifically: increasing urbanization; the current drive to command public expenditure; and the background of dubiousness associated with increasing climate variability under climate change (Guerreiro et al. 2018).

Historically, overflowing risk direction has approached flood hazard prediction in a deterministic fashion (Baldassarre and Montanri 2009). The method used routinely in do uses a deterministic hydraulic (flood) modelling approach that simulates the physical processes decision-making flood flows (Baldassarre et al. 2010). An consequence, characterized by a single set of boundary data, is modelled to produce mapped outputs of the potential depth, velocity and extent of the flood event (Hartanto et al. 2011; Beevers et al. 2012). Withal, the uncertainty due to the dynamic, stochastic and uncertain nature of the climate, hydrological and river processes (Van Vuren 2005) is by and large not considered nor is the underlying model data and choices or the assumptions embedded within the analysis (e.g. return catamenia methodology estimates. Uncertainties in hydraulic (flood) modelling can arise from many sources, and for the purposes of this paper, we follow a structured approach (Warmink et al. 2010; Vicious et al. 2016) to these for hydraulic simulation, classifying them as uncertainties associated with:

the model input: for example, the boundary conditions (e.grand. hydrology and flow estimates) or the underlying bathymetry or topographical information;
the model parameters (e.thou. roughness estimates); or
the choice of the model structure (e.g. the mathematical conception).

Within a alluvion modelling framework, uncertainties cascade through the modelling chain, with each contributing to the variability in the final overflowing extent. Thus, accounting for uncertainties in model prediction this may be accomplished through consideration of multi-model ensembles, perturbed parameter ensembles and/or multi-member (inputs) ensembles. This is often accomplished through Monte Carlo simulation (Mundform et al. 2011) providing estimate solutions to a variety of mathematical problems past performing statistical sampling experiments. Whilst such an arroyo is data intensive and computationally demanding, it is increasingly recognized that in that location is a growing need for outputs in the form of probabilistic overflowing maps (Teng et al. 2017).

With the availability of greater computational power and parallelization of codes, uncertainty assay in inundation model research has become more frequent. Recently, several studies have investigated incorporating uncertainty in models (input and parameters) in hydraulic modelling, flood mapping and alluvion analysis (Apel et al. 2004; 2006; Neal et al. 2013; Savage et al. 2014; Mukolwe et al. 2014; Ali et al. 2015; Teng et al. 2017; Winter et al. 2018; Berends, et al. 2018). The sources of dubiousness considered include inflows, observed data and inundation frequency analysis, model parameters and underlying topographical data. Assessed uncertainties tin be function of a alluvion risk cess and highlight the need to quantify input and parameter variability to prevent unnecessary ecology and economical losses or potential overflowing defence failures (Wobus et al. 2017; Metin et al. 2018; Curran et al. 2019). Some general conclusions have emerged from these studies, suggesting that uncertainty associated with inflow has a significant influence on the flood overflowing prediction. This input uncertainty is more significant than the uncertainty associated with model parameterization (Mukolwe et al. 2014) or topographic data (Ali et al. 2015). However dubiousness associated with climate model parameterization has not been explored.

In regard to the input uncertainties (or inflow), one of the biggest challenges to current inundation risk direction is the uncertainty associated with climate change (Collet et al. 2017) and the touch on it may accept on time to come floods (hazard and exposure) in terms of increasing frequency, duration and magnitude (Collet et al. 2017; 2018a, b; Thober et al. 2018; Visser-Quinn et al. 2019). For example, assuming a not-stationary climate under climate change:

What is the likely magnitude of a future flood outcome (return period) in a particular location, and how may that differ from inundation events at nowadays (Collet et al. 2018a)?
How will this touch the time to come flood footprint of a given event in the hereafter?
Volition this uncertainty exist greater than the inherent uncertainty associated with overflowing frequency interpretation at nowadays (eastward.chiliad. Wintertime et al. 2018; Collet et al. 2017), or will parameter dubiety in the model become more important?

To engagement, no written report has attempted to respond these questions. Previous work by the authors (Collet et al. 2017, 2018a, b) used the Future Flows Hydrology (FFH) database (Prudhomme et al. 2013), a set of spatially coherent daily flow projections, derived from an 11-fellow member Perturbed Parameter Ensemble (PPE) for 282 catchments beyond Great Uk. This work focused on analysing the change in extreme run-off for a range of return period events beyond Cracking Britain. National scale maps were created displaying the doubtfulness related to climate model and probabilistic distributions for flood events from ane in 10 to i in 200 year return periods, with i in 100 and ane in 200 twelvemonth return period events being the IL industry standard for overflowing direction and general planning purposes (east.thou. housing developments). Return periods were determined via 2 extreme value methods, generalized farthermost value (GEV) and generalized Pareto (GP), for two fourth dimension periods, a baseline (1961–1990) and the future (2070–2099). The assay evaluated the incertitude associated with the extreme value (EV) distributions and climate model parameters. For the aforementioned fourth dimension period, the GP and GEV methods saw estimates of run-off with unlike mean values but similar incertitude bounds, whilst between fourth dimension periods, there was an increasing trend in both. Collet et al. (2017, 2018a, b) ended that methods investigating the hereafter flood hazard and extent should account for uncertainty associated with both the climate model parameters and the EV distribution.

Building on this previous work, the aim of this paper is to determine the relative part climate model doubtfulness has in the prediction of future flood extents. To do this, the run-off estimates must be used as hydrograph inputs to numerical hydraulic models to calculate flood inundations (cascading the uncertainty through the modelling process). This is achieved through the application of a probabilistic framework which explicitly captures the dubiety associated with climate model parameterization alongside flood modelling uncertainty (extreme value estimation and hydraulic model roughness specifically).

The potential to reduce computational endeavor using numerical approaches is also considered. The method is demonstrated through application to a Scottish case study for two future return period events (1:ten year and 1:100 year) with focus on the higher magnitude 1:100 twelvemonth event. 2 return period events were considered to address the uncertainty of plumbing equipment extreme value statistics on limited fourth dimension series (xxx year time slices).

Data and models

Instance study

The River Don (Fig. 1a), Aberdeenshire, northward-e Scotland, serves every bit the case study catchment. Information technology has a full catchment area of approximately 1300km², and the river flows from the Cairngorms National Park to the Due north Ocean at Aberdeen. Previous work by the authors (Collet et al. 2017, 2018a) identified the River Don as a catchment likely to encounter an increased flood hazard as a outcome of climate change.

This piece of work focuses on a 5 km achieve of the river downstream of the Parkhill catamenia gauge (station number: 11001, NRFA), which flows through Dyce, a suburb of Aberdeen. The gauging station has flow records since 1950 and shows frequent inundation events (almost recently in 2016); the highest water levels on tape occurred on 8 January 2016. This is a relatively short and hydraulically simple reach of the river, with a topographically variable floodplain, allowing the exploration of complex uncertainties in a systematic fashion.

Hydraulic model: LISFLOOD-FP

LISFLOOD-FP (Bates and De Roo 2003) is selected as the hydraulic model, due to its computational efficiency, and reduced physics formulation. The LISFLOOD-FP model requires input data in the form of boundary weather, a digital elevation model, channel geometry and friction coefficients. The upstream purlieus was located at the Parkhill gauging station, with a free boundary located v km downstream of the estimate equally the outflow of the domain (Fig. 1). The Scottish Environment Protection Agency (SEPA 2015) provided the digital peak model (1 chiliad LiDAR) as well as thirteen river cross sections from which aqueduct geometry was derived. A land cover map was determined using OS data at a 5 chiliad resolution, with the associated Manning'south due north determined from the literature (Chow 1959); see Table one.

Tabular array 1 Roughness coefficients, an input to the hydraulic model LISFLOOD-FP, for the study expanse

Full size table

The LISFLOOD-FP model was congenital using the sub-filigree aqueduct solver for main channel flow and the acceleration solver for floodplain menstruation. The model was calibrated for November 2002, October 2002 and September 1995 inundation events [daily gauged flow and digitized maximum inundation extent maps (SEPA)] and validated to National Flood Run a risk Maps for i:10, 1:200 and ane:one thousand return periods. Roughness (Manning's n) was adjusted in order to maximize agreement between observed and simulated maximum alluvion extents using the fit statistic Fstat (the overlap between observed and predicted as a proportion of moisture cells); Fstat ranges from 0 to 1, with 1 indicating perfect understanding (Goudet 1995).

Period projections

Following Collet et al. (2017), this report utilized daily flow projections from the FFH database, an 11-member perturbed parameter ensemble. It is through this perturbing of parameters that climate model parameter dubiety may be explored. Flow projections were extracted for the Don River at Parkhill gauging station. Every bit is standard in climate impact assessment (Allen et al. 2018), this report focuses on two thirty-twelvemonth climate normals, a 1961–1990 baseline and the 2080s (2069–2098). To ensure consistency in the analysis, merely flow projections are utilized. The focus is on the change in extreme events between the time periods, i.e. the changes in height period and doubtfulness due to climate modify.

Methods

This newspaper aims to understand the role of climate model parameter uncertainty in futurity alluvion inundation projections through a instance study catchment (Sect. two.one). In club to reach this aim, the office of uncertainty associated with climate model parameters was explored in context with other sources of incertitude by sampling across the distributions of inputs for the hydraulic model following four structured tests (Sect. iii.1 and Fig. 2):

a.
Climate model uncertainty (CMU), as quantified in Collet et al. (2017) (Fig. 2a). The empirical CDF of the return period estimates for the 11-member ensemble;
b.
Extreme value (structure and parameter) doubt (EVDU) for ii different extreme value models, the generalized farthermost value (GEV) and generalized Pareto (GP) distributions (Fig. 2b). Parameter uncertainty is considered through the 95% confidence intervals. After Collet et al. (2017), in order to isolate this uncertainty from the climate model incertitude, only the ensemble median return menstruum judge was considered;
c.
Hydraulic model parameter uncertainty (HMPU; Fig. 2c), the roughness (Manning's n) values for the floodplain and channel definition. This test fixes the Q _peak used to the median estimate across the ensemble;
d.
Total incertitude (Fig. second), to determine the relative dubiety associated with each source. Full doubtfulness is measured equally a combination of tests a–c, where all iii factors are varied.

The tests were applied for the 1:ten- and 1:100-yr return period flood events for the baseline (1961–1990) and future (2070–2099) time periods (to understand the influence of a non-stationary climate), beyond the 11 every bit plausible hydroclimatological realizations for the Parkhill gauge station.

The tests were applied through two probabilistic frameworks (Sect. 3.ii). The first, a full Monte Carlo analysis, is a computationally expensive approach where a big number of randomly sampled realizations are assumed to represent the distributions in their entirety. Whilst computational power is increasing, there is nevertheless a need to explore more time-efficient strategies which may result in a similar estimate of the full uncertainty. As a first stride, this paper as well explored a simple stratified sampling arroyo in guild to reduce this computational burden.

Incertitude tests

Climate model parameter uncertainty (CMU)

In order to determine the 1:10- and 1:100-year return period inundation events (Q _peak), GEV and GP distributions (Coles 2001) were fitted to annual maxima (AMAX) and peak-over-threshold (POT) flows, respectively. Come across Collet et al. (2017) for the detailed methodology. For both return periods (ane:10 and 1:100) and each extreme value model (GEV and GP), Q _pinnacle values were computed for the equally likely ensemble members, per time period ($N=11\times ii$). Using the empirical CDF constructed by the eleven values of $N$, a total of north Q _peak values were randomly sampled across the distribution (where due north is adamant past the probabilistic framework outlined in Sect. 3.2, i.due east. full Monte Carlo or LHS). These $n$ Q _elevation values served as the input to the hydraulic model (see Fig. 2a). The input hydrograph for the LISFLOOD-FP model was scaled for each simulation using the sampled Q _summit, generating corresponding hydrographs that were used to run the hydraulic model with the spatially distributed roughness (Manning'south north) values held constant.

Extreme value model (structure and parameter) doubt (EVDU)

For the N ensemble members and time periods, the median Q _peak value was determined across both EV distributions, after Collet et al. (2017). This median represents the structural uncertainty, whilst 95% confidence intervals capture parameter uncertainty. This range was randomly sampled following a normal distribution, generating n hydrographs which stand for the boundary weather condition in the hydraulic model. The spatially distributed roughness (Manning's north) values were held constant in the hydraulic model (meet Fig. 2b).

Hydraulic model parameter uncertainties (HMPU)

Assuming a uniform distribution, the HMPU was determined past sampling across the Manning'south n parameter space (range—see Table one), producing due north maps. The input hydrograph to the hydraulic model was held constant, corresponding to the median value of the GEV distribution across the 11 ensemble members (see Fig. 2c).

Full dubiety

Full dubiousness simulations were then completed where all iii factors from tests a–c were varied: (1) CMU, the uncertainty in Q _elevation from across the empirical CDF; (2) EVDU, where the median Q _peak and 95% confidence intervals represent the structure and parameter uncertainty, respectively; and (3) the HMPU, captured through maps of roughness value.

Probabilistic frameworks

Full probabilistic approach: Monte Carlo analysis

The total probabilistic arroyo adopts a Monte Carlo arroyo (Mundform et al. 2011). For each uncertainty test (Sect. iii.ane), factor(s) are varied by sampling across the respective range of values with an associated probability distribution. A pseudo-random number generator (PRNG) was used to sample (due north = 10,000) from the prescribed distributions for each test, for a full of 80,000 simulations ($2\;{\text{fourth dimension}}\;{\text{periods}} \times iv\;{\text{tests}} \times due north$).

Reducing the computational burden: stratified sampling: Latin hypercube sampling

Latin hypercube sampling (LHS) divides the distributions into regions of equal probability and randomly samples a value from each region. This increases the sampling efficiency past ensuring compatible coverage of the parameter space whilst using fewer samples (Helton and Davis 2003). Theoretical studies have shown that the sampling error for LHS, $O\left(\frac{ane}{\sqrt{northward}}\right)$, is quadratically faster than randomly sampled Monte Carlo simulations, $\left(\frac{1}{due north}\right)$; simply put, LHS can accomplish the same accuracy as n^two Monte Carlo simulations (Aistleitner et al. 2012). However, in practice, as the dimensionality of the input increases, the accuracy reduces (Huntington and Lyrintzist 1998). An extension of LHS, orthogonal sampling, has been employed in this study every bit this ensures that each subspace is evenly sampled.

The distributions for each test were binned, $northward=\{50, 125, 250, 500, 1000\}$. Each bin was sampled once, with x replicates performed to test sample size variability. These replicates were tested to determine the minimum number of samples required for convergence (hateful and standard deviation of the flooded area in the full Monte Carlo analysis).

Results

Model calibration

The final calibrated model has a reasonable fit and was found to perform all-time for college (out of banking concern) events, making it suitable for this study (Tabular array 2), where significant flow events (1: 10-year and i:100-yr return flow) are chosen for analysis. Water levels at specified cross sections were inside ± 0.2 thou. The calibrated model was validated against the more farthermost out-of-bank events (1:200-year and i:thou-year events), for which information were available, performing well and confirming reasonable calibration of the model.

Table 2 Hydraulic model calibration and validation results: fit between observed and modelled extent

Total size tabular array

Probabilistic framework: total Monte Carlo analysis

The outcomes from the full Monte Carlo analysis are start considered. The ane:100-twelvemonth return menstruation event results are presented in the main paper, whilst the results for the one:10-year return period event are presented in the supplementary material. For the 1:100-year render period event, Fig. iii depicts the flooded expanse versus Q _peak per test; the test results are aggregated in Fig. 4, allowing comparing of the baseline and future uncertainties. Modify in flooded area, per test, is farther considered in Fig. 5 through a series of frequency distributions. The change in output hateful values (Delta mean) is indicated for each uncertainty, further measures of change are provided in Table iii. Finally, Fig. 6 presents the probabilistic flood maps for the total uncertainty (test d); see supplementary data for the probabilistic flood maps for tests a–c. The results consider the influence the unlike sources of uncertainty provide, before exploring the impacts of because all sources together. Results are presented in supplementary material for the 1:10-year return menstruum effect, equally shown in Figs. ten, 11 and 12.

Table 3 Percentage change to serial statistics from the baseline to the future for a 1:100 year return catamenia

Total size table

Sources of uncertainty: tests a–c

Figures iii and 5 and Table three consider each source of doubtfulness, as well every bit the total uncertainty, separately (ane:100-yr render menstruation event). It tin be seen that, in the case of climate model parameter dubiousness (a), the range of incertitude is smallest on the baseline, with ∆-mean increasing past twenty% in the future (Fig. 5; bias between the samples). With a ∆range of almost 40% (Table 3), the overall climate model parameter incertitude can be seen to increment. Like changes in mean (19%, Table iii) are observed for the EV model structure and parameter doubtfulness (b). A change in distribution from the baseline (bimodal) to the future (normal) can exist observed in Fig. 5b. This is likely due to topographic thresholds present in the baseline simulations but not observed in future outputs due to water levels higher than these features. Interestingly, the results of the 1:10-year return period event (supplementary textile) reinforce the importance of the topographic control in this case study.

Despite this, the increase in overall uncertainty is lower than for the climate model parameter (∆range = 24%, Table 3). Test c saw Q _pinnacle held abiding as the roughness maps were varied; accordingly, the plot in Fig. 3c is a horizontal line. The change in roughness under test c presents a different flick. The ∆-mean (31%) is higher than in tests a and b, whilst the ∆-range sees a big reduction (64%). Overall, hydraulic model parameter uncertainty contributes the least to the range of modelled flooded area for both the baseline and hereafter time periods, whilst the source of uncertainty contributing the most to the range of output is marginally EVD (over CMU). This finding is irrespective of render period explored (1:10 vs. 1:100), an interesting upshot peculiarly given the assumption that EVD dubiousness increases with higher return flow estimates. This would announced to agree with previous studies, e.g. Apel et al. (2018).

Varying the roughness (test c) results in a range of flood extents across similar Q _peak values (roughness influencing the h2o level and thus extent). This adds a width or fuzziness to the flooded area ~Q _height human relationship (Fig. 4a, b) and results from the consideration of two different factors (dimensions) within the dubiety analysis.

It is interesting to notice from Fig. 3 that, contained of the exam, inflexion points be in the Q _peak ~ flooded area relationship, where the relationship between these variables changes. These points are observed throughout all iv tests (for both baseline and hereafter) and for both return period simulation ensembles, at effectually Q _superlative = 230 m³/s and around Q _peak = 270 grand³/southward. The relationship with flooded surface area flattens out between these two Q _peak values (230–270 thou³/s), indicating that the flooded area is increasing more rapidly with an increasing Q _peak. Information technology is very likely that these inflexions are related to topographic controls in the case written report expanse. These observed relationships are context specific, and the details are non-transferable betwixt case studies.

Total uncertainty—test d

Test d (Figs. 3d, 5d) explored the total uncertainty through consideration of the full parameter infinite. From Tabular array 3, it can be seen that ∆mean and ∆range are within the range of the individual sources of doubtfulness (tests a, b and c). Two dimensions of dubiety, the hydraulic model parameter (HMPU) and the input uncertainty (Q _peak), arising from both climate model and farthermost value distribution uncertainties (CMU and EVDU) are shown to result in fuzziness in the relationship. Figure 6 shows clearly the affect on the overflowing area and the alter in probability of occurrence from the baseline to the hereafter fourth dimension periods. These probabilistic flood maps prove areas which get much more probable to be flooded in the future given the aforementioned return period issue, besides equally some areas which are newly inundated in the hereafter.

Probabilistic framework: stratified sampling LHS results

Effigy 7 shows the results of the convergence of the hateful and standard deviation for the LHS written report for the iv tests over 2 time periods (baseline and future). In the kickoff three tests, where the sources of doubtfulness are considered separately, the sample mean converges for due north = 1000, thereby suggesting that the northward = 1,00 runs are able to capture the total dubiousness from the Monte Carlo assay (n = ten,000). This is seen clearly on the baseline where the uncertainty range in the inputs is lower than for the future. For the hydraulic model parameters, the uncertainty range is relatively pocket-size; despite this, the uncertainty is well captured. By comparison, when the total dubiety is explored using the full parameter space, the LHS stratified sampling performs poorly. This is probable due to the increase in dimensionality (three components, collapsed into two parameters) influencing the efficacy of stratified sampling; this is further considered in the discussion.

A total of ten ensembles were considered for each $n = \{ fifty,125,250,500,1,000\}$ sample size. For the full uncertainty, the frequency distribution for each sample size is presented for the baseline (Fig. 8) and future (Fig. ix). For the baseline (Fig. 8), it tin exist seen that the shape of the distribution (bimodal) is well replicated from n = 125; even so, with loftier variability across the ensemble, the associated incertitude is high. At n = 1000, the variability is significantly reduced and in the variability associated with the ensemble, ability to replicate the shape has reduced. For the hereafter time period, this reduction in variability is more distinct given the greater range of flooded area. In the future (Fig. 9), the sample size required for suitable replication of the frequency distribution increases to a minimum of n = 250. Improvements in the replication of the tails of the distribution (900,000–1,000,000 chiliad²) tin exist seen at n = 500, whilst the ensemble can be seen to move towards convergence at n = 1000. From Figs. viii and nine, it is articulate that the LHS stratified sampling is able to provide a reasonable representation of the full Monte Carlo analysis. Notwithstanding, information technology is likewise apparent that n = 100 is insufficient to capture the full range of uncertainty (equally n ² MC simulations; Aistleitner et al. 2012).

Discussion

Main results: uncertainty sources

For the full Monte Carlo simulations (for both return period events), the results for the baseline and future advise that the dominant source of uncertainty is the estimated input hydrograph. This is a reflection of both the uncertainties associated with the parameterization of the regional climate model (CMU) and the extreme value method (EVDU). These findings are consequent for both return periods tested and with recent studies by Mukolwe et al. (2014) and Ali et al. (2015). It can thus be inferred that the relative scale of input hydrograph dubiousness to hydraulic model parameter doubtfulness (CMU and EVDU > HMPU) is transferable between sites. However, the specifics of the relationship between Q _summit and flooded area are location and context specific, with local topographic controls, such as floodplains controlling the processes; thus, each case study location will evangelize dissimilar findings. In such cases, where wetting and drying of cells becomes critical and therefore more sensitive to the modelling process (parameterization, technique, mathematical formulation), the relationship may become more than complex.

The results presented in Figs. three and four suggest that there are certain inflexions within the Q _meridian and flooded expanse human relationship, where the human relationship changes (230 m^three/s and 270 m³/s). Here, there is a range of values of Q _peak for which flooded area extent does not vary much despite a relatively larger increase in Q _peak. The identification of these thresholds may be possible with fewer simulations depending on the complexity of the modelled stretch; however, probabilistic methods are required to fully represent the relationship and provide useful insights for planners and consultants. For case, in the Great britain this could inform the specification of inundation zones in policy guidance from the Environment Bureau, Natural Resources Wales or the Scottish Environmental Protection Agency.

The quantification of this human relationship may permit for the utilise of surrogate models. Surrogate modelling is a technique that approximates a more than computationally expensive model whilst honouring the underlying physics. A small sample size of well-chosen information points is used to model the response of the original model, creating an emulator that requires fewer simulations to quantify the uncertainty. Surrogate models accept already demonstrated their applicability to hydrological uncertainty quantification (Roy et al. 2018; Moreno-Rodenas et al. 2018) with the most mutual surrogate models being polynomial chaos expansion (PCE) and Gaussian processes (GP). The reduced sample size allows for more numerically intensive models to be used with little increase in computational burden, producing more accurate and efficient flood hazard assessments.

Non-stationarity in climatic change impacts to flood inundation extents.

Climate ensembles project changes in climatic variables such as temperature and atmospheric precipitation (IPCC 2014), the two main drivers of the hydrological bicycle. Inquiry into the effects of climatic change on the hydrological system has been ongoing in excess of twenty years (Augustin et al. 2008; Christierson et al. 2012). Although highly uncertain, recent work in United kingdom of great britain and northern ireland and the UK indicates an important change in future extreme flows (Collet et al. 2017; 2018a, b); spatial variability has also been highlighted as an important factor (Thober et al. 2018; Visser-Quinn et al. 2019).

To sympathise the affect on population, exposure and subsequent risk, there is a articulate need to characterize and quantify this doubt, cascade this into flood inundation assessment using flood models and compare consistently between the nowadays climate (baseline) and in the future (e.g. the 2080s) climate model estimates. The results of this study bear witness that input hydrographs boss the uncertainty in flood modelling and interestingly showed that uncertainty associated with climate model parameterization is at least equally important to capture as the incertitude associated with the extreme value distribution fit, irrespective of return period estimate. The range of possible inundation extents which capture the uncertainty in input parameters is significant for both return periods on the baseline and time to come, which suggests that probabilistic alluvion maps take an important role in future flood inundation assessments.

Presently, the UK utilizes standard hard edge flood maps. Derivation of probabilistic alluvion maps (such equally Fig. 6) for a 1:100-year return period inundation extent could serve to illustrate the projected change in flood probability and thus exposure through a model domain. Whilst this is computationally expensive and may be more hard to communicate, probabilistic maps tin can provide a new level of data for use in time to come adaptation planning. Work over the last decade (Baldassarre and Montanri 2009) has increasingly chosen for this modify in the way flood mapping is considered. Recent studies (Teng et al. 2017) predict that, going forward, probabilistic mapping will be in greater demand equally the industry, insurance companies and policy makers alike grapple with how best to capture and convey dubiety in projections within the context of flood legislation such as the EU Floods Directive (binding legislation for Eu members states).

With an increase in mean flooded area for full uncertainty of between 18 and 22% (based on the two render periods explored in this newspaper), this work shows that it is critical to capture the modify to inflows as a result of a non-stationary climate in futurity inundation assessments and admit the uncertainty associated with its projection. Thus, it seems sensible to suggest the non-stationarity of climate and its influence on extreme events in the future; aslope the uncertainty associated with these projections needs to be captured more routinely and explicitly in flood chance assessments, through probabilistic flood hazard maps.

Reducing the computational cost of uncertainty studies

This report has highlighted the ability of LHS to reduce computational price with minimal to no loss of accuracy. Figure 7 shows that, for each source of uncertainty considered, a sample size of north = 500 was able to converge to the mean and standard difference of the full Monte Carlo simulation. For the total dubiety, where both Manning's northward and the input hydrograph are varied, a sample size of due north = 1000 was required (Figs. 8 and 9) to highlight variability in the LHS ensemble, suggesting n = 1000 provides the amend replication of the total uncertainty. This was shown to exist true for non-normal distributions (bimodal) in particular, with thresholds in the topography influencing tipping points (i.east. either wet or dry).

The theory indicated that LHS is quadratically faster, achieving the aforementioned accuracy every bit due north² Monte Carlo simulations (Aistleitner et al. 2012). The empirical findings in this study suggest that in order to achieve adequate replication of the mean, standard deviation, range and frequency of the flooded area distribution, a sample size of n = g is required (i.east. ${\left(\frac{north}{10}\right)}^{2}$). This is still computationally intensive, and the added complication is the dimensionality of the uncertainty parameters considered in this report. Therefore, there is a critical demand to investigate ways in which the computational toll can exist reduced. In this written report, a reduced physics numerical engine (LISFLOOD-FP) was used to drive the assay. The benefit of this approach is the speed of numerical calculation. Whilst reducing the number of runs by a gene of x is useful, there is a articulate need for farther work. However, in manufacture the majority of inundation models utilize more computationally intensive fully 2nd flood models (e.grand. TuFLOW, TELEMAC, MIKE21) which are non well suited to this probabilistic arroyo. Developments in other fields may provide alternative solutions. One case is Markov chain Monte Carlo (MCMC), an dubiousness quantification method that uses Markov concatenation to sample from the input distribution. The Markov chain is substantially a random walk through the parameter space with each step depending simply upon the previous value (memoryless). The well-nigh common MCMC method is the Urban center–Hastings algorithm (Chib and Greenberg 1995) which requires a reduced number of input values compared to full Monte Carlo. The MCMC has been shown to converge faster than Monte Carlo (Elsakout et al. 2015). With this increased charge per unit of convergence, the application of more complex hydraulic models requiring larger simulation times is more than feasible. The computational cost may exist further reduced through awarding of Multi-Level Markov Concatenation Monte Carlo (MLMCMC), where multiple grids are used to judge quantities of involvement, whilst applying a Markoff chain approach to sampling to ensure computational efficiency (Elsakout et al. 2015).

Limitations of the study

In that location are some limitations which are worth noting for this written report and which warrant farther consideration in future research. This study utilized data from the FFH database for a single emissions scenario, SRES A1B. This data is derived from the UKCP09 projections, a perturbed physical ensemble based on a unmarried climate model (HADCM3). Projections from a hydroclimatological modelling concatenation that utilizes a multi-model ensemble may yield different results (e.chiliad. the Border projections, Thober et al. 2018). Other important considerations for future work include:

Modelling with the near recent CMIP5 projections;
Comparison across multiple emissions scenarios (e.g. RCP2.half dozen and RCP8.v);
Modelling across a range of topographical diverse catchments or longer river reaches, which explores the dynamic interaction of floodplain inundation, inundation defence failures and flood management within river reaches—adding hydraulic complexity into the assay.

This is particularly relevant for studies looking at surrogate methods (Sect. 5.1) and culling uncertainty quantification methods (Sect. 5.three). Whilst this study demonstrated a reasonable reduction in computational cost through the application of a LHS arroyo, fully 2d codes, as used in industry, require more than a tenfold reduction in simulations. Alternative approaches need to be explored in club to open up the possibility of probabilistic inundation assessments every bit function of routine practice, particularly in industry.

Conclusions

This report has explicitly captured the dubiousness associated with climate model parameterization and the flood modelling process (extreme value estimation and hydraulic model roughness specifically) using a probabilistic framework and investigated the influence of climatic change projections on hereafter flood run a risk predictions. The findings clearly demonstrate the need to capture climatic change projections in futurity overflowing inundation interpretation. The results suggest that the uncertainties associated with hydrological uncertainty, or the menstruation input, result in the nearly substantial doubt in the flood alluvion extent. Certainly, hydraulic parameter uncertainty is considerably less influential on the spread of inundation extents. These findings are consistent with contempo studies (Mukolwe et al. 2014; Ali et al. 2015).

Doubtfulness in the inflow parameters is important to the eventual consequence (inundation extents), and thus, there is a need to routinely include this doubt into flood inundation studies. From the findings of this report, the uncertainty associated with farthermost value distribution models can outcome in a reasonably similar range of inundation extents as the dubiety arising from climate model parameterization; however, more importantly non-stationarity in climate (i.eastward. the impact of climate alter) results in a considerable shift in the mean of flood extents, too as the range. Thus, alongside capturing the doubt in hydrological inputs and cascading this through hydraulic models, there is an urgent need to include not-stationarity in the climate.

Tipping points in the relationships betwixt the pinnacle flows and the resulting flood extents can exist useful for policy and in practice (e.g. planning control); all the same, these will exist very context specific. These tipping points are strongly controlled past topographic features in the flood written report areas, and this study constrained this issue through the choice of a curt, reasonably uncomplicated hydraulic reach. For case, big floodplains will influence the relationship betwixt alluvion peak and overflowing area. Further investigation of such relationships is needed through more than case report assessments.

Finally, in practice the power to run repeated hydraulic simulations to capture the full uncertainty in the input and parameter space is unlikely. Due to computational burden, capturing the full uncertainty is probable to be a luxury afforded in the research surroundings simply. Thus, there is a real demand to investigate the opportunities to reduce computational brunt by reducing the number of hydraulic model runs necessary whilst withal capturing the total range on possible inundation extents. This paper has explored a elementary first arroyo to reducing simulations through the use of a stratified sampling approach (LHS). This method, whilst found to be effective, still has not reduced the number of runs sufficiently for it to be a routinely viable pick in practice. Therefore, more sophisticated methods are needed to quantify uncertainty when considering uncertain parameter space in more than one dimension. This paper has suggested several unlike methods that could be explored, by looking to other disciplines for inspiration, and is recommended as an avenue for future research.

References

Aistleitner C, Hofer M, Tichy R (2012) A central limit theorem for Latin hypercube sampling with dependence and application to exotic basket option pricing. Int J Theor Appl Finance 15(7):1250046. https://doi.org/10.1142/S021902491250046X
Ali MA, Solomatine D, Baldassarre Chiliad (2015) Assessing the bear upon of different sources of topographic data on i-D hydraulic modelling of floods. Hydrol Earth Syst Sci 19:631–643. https://doi.org/10.5194/hess-xix-631-2015

Commodity Google Scholar
Allen MR, Dube OP, Solecki W, Aragón-Durand F, Cramer W, Humphreys Southward, Kainuma K, Kala J, Mahowald North, Mulugetta Y, Perez R, Wairiu Yard, Zickfeld K (2018) Framing and context. In: Masson-Delmotte V, Zhai P, Pörtner H-O, Roberts D, Skea J, Shukla PR, Pirani A, Moufouma-Okia West, Péan C, Pidcock R, Connors S, Matthews JBR, Chen Y, Zhou X, Gomis MI, Lonnoy E, Maycock T, Tignor G, Waterfield T (eds) Global warming of 1.5 °C. An IPCC special report on the impacts of global warming of i.five °C above pre-industrial levels and related global greenhouse gas emission pathways, in the context of strengthening the global response to the threat of climatic change, sustainable development, and efforts to eradicate poverty
Apel H, Thieken AH, Merz B, Blöschl G (2004) Inundation risk assessment and associated dubiousness. Nat Hazards Earth Syst Sci 4:295–308

Article Google Scholar
Apel H, Thieken AH, Merz B, Blöschl Thou (2006) A probabilistic modelling organisation for assessing inundation risks. Nat Hazards 38:79–100. https://doi.org/10.1007/s11069-005-8603-7

Commodity Google Scholar
Augustin N, Beevers L, Sloan W (2008) Predicting river flows for future climates using an autoregressive multinomial logit model'. Water Resour Res. https://doi.org/ten.1029/2006WR005127

Article Google Scholar
Bates PD, De Roo APJ (2000) A simple raster-based model for flood inundation simulation. J Hydrol 236:54–77

Commodity Google Scholar
Baldassarre G, Montanri A (2009) Dubiety in river discharge observations: a quantitative analysis Hydrol. Globe Syst Sci xiii:913–921

Article Google Scholar
Baldassarre 1000, Schumann Yard, Bates P, Freer J, Beven M (2010) Flood-plain mapping: a critical discussion of deterministic and probabilistic approaches. Hydrol Sci J 55(3):364–376

Commodity Google Scholar
Balica SF, Beevers L, Popescu I, Wright NG (2013) Parametric and physically based modelling techniques for flood risk and vulnerability cess: a comparison. J Environ Model Softw 41(3):81–92. https://doi.org/x.1016/j.envsoft.2012.11.002

Commodity Google Scholar
Beevers 50, Douven W, Lazuardi H, Verheij H (2012) Cumulative impacts of road developments in floodplains. Transp Res Part D Transp Environ 17(5):398–404

Article Google Scholar
Beevers L, Walker G, Strathie A (2016) A systems arroyo to inundation vulnerability. Ceremonious Eng Environ Syst 33(3):199–213. https://doi.org/ten.1080/10286608.2016.1202931

Article Google Scholar
Berends K, Warmink JJ, Hulscher SJMH (2018) Efficient dubiousness quantification for bear upon assay of man interventions in rivers. Environ Model Softw 107:50–58. https://doi.org/10.1016/j.envsoft.2018.05.021

Commodity Google Scholar
Chib S, Greenberg E (1995) Understanding the urban center–hastings algorithm. Am Stat 49(4):327–335. https://doi.org/10.2307/2684568

Article Google Scholar
Chow VT (1959) Open channel hydraulics. McGraw-Hill, New York

Google Scholar
Christierson BV, Vidal JP, Wade SD (2012) Using UKCP09 probabilistic climate information for United kingdom of great britain and northern ireland water resource planning. J Hydrol 425:48–67. https://doi.org/10.1016/j.jhydrol.2011.12.020

Article Google Scholar
Coles S (2001) An Introduction To Statistical Modelling Of Extreme Values. Springer, London

Book Google Scholar
Collet 50, Beevers Fifty, Prudhomme C (2017) Assessing the impact of climate alter on extreme flows across Great Great britain. Water nine(2):103. https://doi.org/x.3390/w9020103

Article Google Scholar
Collet L, Beevers L, Stewart M (2018a) Decision-making and flood risk incertitude: statistical dataset analysis for alluvion risk cess. Water Resour Res. https://doi.org/10.1029/2017WR022024

Article Google Scholar
Collet L, Formetta G, Harrigan S, Prudhomme C, Beevers L (2018b) Future hot-spots for hydro-hazards in Corking Uk: a probabilistic assessment. Hydrol Earth Syst Sci 22(10):5387–5401

Article Google Scholar
Curran A, de Bruijn KM, Klerk WJ, Kok Thou (2019) Big scale inundation take chances analysis past including defence failures on the Dutch river system. Water eleven:1732. https://doi.org/10.3390/w11081732

Commodity Google Scholar
Elsakout DMA, Christie MA, Lord GJ (2015) Multilevel Markov Chain Monte Carlo (MLMCMC) for uncertainty quantification. In: SPE Due north Africa technical conference and exhibition, xiv–xvi September, Cairo, Egypt Society of Petroleum Engineers
Goudet J (1995) FSTAT (version i.2): a computer program to calculate F-statistics. J Hered 86:485–486

Article Google Scholar
Guha-Sapir D (2018) EM-DAT: The Emergency Events Database - Universite catholique de Louvain (UCL) - CRED, - www.emdat.exist, Brussels, Belgium (Accessed: 22/eleven/2018).
Guerreiro S, Dawson RJ, Kilsby C, Lewis Due east, Ford A (2018) Futurity heat-waves, droughts and floods in 571 European cities. Environ Res Lett. https://doi.org/x.1088/1748-9326/aaaad3

Article Google Scholar
Hartanto IM, Beevers L, Popescu I, Wright NG (2011) Application of a littoral modelling code in fluvial environments. Environ Model Softw 26(12):1685–1695

Commodity Google Scholar
Helton JF, Davis J (2003) Latin hypercube sampling and the propagation of incertitude in analyses of circuitous systems. Reliab Eng Syst Saf 81:23–69

Commodity Google Scholar
Huntington DC, Lyrintzist S (1998) Improvements to and limitations of Latin hypercube sampling. Prob Engng Mech thirteen(4):245–253

Article Google Scholar
IPCC (2014) Climatic change 2014: synthesis written report. In: Pachauri Thousand, Meyer LA (eds) Contribution of working groups I, 2 and III to the fifth cess report of the intergovernmental panel on climate change [core writing team]. IPCC, Geneva, Switzerland
Jonkman S, Vrijling J (2008) Loss of life due to floods. J Flood Adventure Manag 1:43–56

Commodity Google Scholar
Metin AD, Dung NV, Schröter K, Guse B, Apel H, Kreibich H, Vorogushyn Due south, Merz B (2018) How do changes along the run a risk chain affect flood risk? Nat Hazards Earth Syst Sci 18(3089–3108):2018. https://doi.org/10.5194/nhess-xviii-3089-2018

Article Google Scholar
Moreno-Rodenas AM, Bellos V, Langeveld JG, Clemens FHLM (2018) A dynamic emulator for physically based flow simulators under varying rainfall and parametric conditions. H2o Res 142:512–527

Commodity Google Scholar
Mukolwe MM, Di Baldassarre G, Werner MGF, Solomatine DP (2014) Flood modelling: parameterisation and inflow doubtfulness. In: Proceedings of the Ice—water management, vol 167, pp 51–60
Mundform DJ, Schaffer J, Kim M-J, Shaw D, Thongteeraparp A, Supawan P (2011) Number of replications required in Monte Carlo simulation studies: a synthesis of iv studies. J Modernistic Appl Stat Methods 10(1):iv. https://doi.org/x.22237/jmasm/1304222580

Article Google Scholar
Neal J, Keef C, Bates P, Beven K, Leedal D (2013) Probabilistic flood risk mapping including spatial dependence. Hydrol Process 27:1349–1363

Article Google Scholar
Prudhomme C, Haxton T, Crooks S, Jackson C, Barkwith A, Williamson J, Kelvin J, Mackay J, Wang L, Young A, Watts G (2013) Future Flows Hydrology: an ensemble of daily river flow and monthly groundwater levels for utilize for climate change bear upon assessment across Great Britain. Earth Syst Sci Information 5:101–107. https://doi.org/10.5194/essd-5-101-2013

Article Google Scholar
Roy NPT, El Moçayd A, Ricci S, Jouhaud N, Goutal One thousand, De Lozzo M, Rochoux C (2018) Comparison of polynomial chaos and Gaussian process surrogates for incertitude quantification and correlation estimation of spatially distributed open-channel steady flows. Stoch Env Res Gamble Appraise 32(6):1723

Article Google Scholar
Savage J, Bates P, Freer J, Neal J, Aronica K (2014) The impact of calibration on probabilistic flood inundation maps using a 2D hydraulic model with uncertain boundary conditions. In: Vulnerability, uncertainty, and run a risk, pp 279–289
Savage J, Bates P, Freer J, Neal J, Aronica GT (2016) When does spatial resolution become spurious in probabilistic flood inundation predictions? Hydrol Procedure 30(xiii):2014–2032. https://doi.org/10.1002/hyp.10749

Commodity Google Scholar
Sayers Pb, Horritt Thousand, Penning-Rowsell E, McKenzie A (2015) Climatic change risk assessment 2017: projections of hereafter flood run a risk in the United kingdom of great britain and northern ireland. Enquiry undertaken by Sayers and Partners on behalf of the Committee on Climate Change. Published by Committee on Climate Change, London
SEPA (2015) Provision of numerical model. Personal advice
Thober S, Kumar R, Wanders N, Marx A, Pan M, Rakovec O, Samaniego Fifty, Sheffield J, Wood EF, Zink M (2018) Multi-model ensemble projections of European river floods and loftier flows at i.5, 2, and 3 degrees global warming. Environ Res Lett 13(i):014003

Commodity Google Scholar
Teng JAJ, Jakeman J, Vaze BFW, Croke D, Dutta Southward, Kim southward(2017) Flood inundation modelling: A review of methods, contempo advances and dubiousness analysis, Environmental Modelling & Software 90:201–216, https://doi.org/ten.1016/j.envsoft.2017.01.006.
Van Vuren S (2005) Stochastic modelling of river morphodynamic. PhD thesis, Delft Academy of Technology, IOS Press, Amsterdam, Netherlands. ISBN: 90-407-2605-1
Visser-Quinn AG, Beevers Fifty, Collet Fifty, Formetta G, Smith G, Wanders N, Thober S, Pan M, Kumar R (2019) Spatio-temporal assay of compound hydro-hazard extremes across the UK. Adv Water Resour. https://doi.org/10.1016/j.advwatres.2019.05.019

Article Google Scholar
Warmink J, Janssen J, Booij M, Krol Thou (2010) Identification and classification of uncertainties in the application of environmental models. Environ Model Softw 25(12):1518–1527

Article Google Scholar
Winter B, Schneeberger K, Huttenlau M, Stotter J (2018) Sources of incertitude in a probabilistic flood risk model. Nat Hazards 91:431. https://doi.org/x.1007/s11069-017-3135-5

Commodity Google Scholar
Wobus C, Gutmann East, Jones R, Rissing M, Mizukami N, Lorie M, Mahoney H, Wood AW, Mills D, Martinich J (2017) Climate change impacts on inundation risk and asset damages within mapped 100-year floodplains of the contiguous United States. Nat Hazards Earth Syst Sci 17(2199–2211):2017. https://doi.org/10.5194/nhess-17-2199-2017

Article Google Scholar

Download references

Acknowledgements

The Future Menstruation Hydrology dataset is available from the Heart for Ecology and Hydrology website at https://www.ceh.ac.uk/services/future-flows-maps-and-datasets. Flood records and topographic data were bachelor on license agreement with the Scottish Ecology Protection Agency (SEPA). This work was carried out as function of the EPSRC EP/L026538/1 project 'Accounting for Climatic change Uncertainty in Flood Hazard Prediction' and EPSRC EP/N030419/i 'Water Resilient Cities: climate uncertainty and urban vulnerability to hydrohazards'

Author information

Affiliations

Water Resilient Cities Grouping, Institute for Infrastructure and the Environment, Heriot-Watt University, Edinburgh, U.k.

Lindsay Beevers, Lila Collet, Gordon Aitken, Claire Maravat & Annie Visser
Irstea, HYCAR Research Unit, 1 rue Pierre Gilles de Gennes, 92 160, Antony, France

Lila Collet

Respective writer

Correspondence to Lindsay Beevers.

Boosted information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary cloth

Rights and permissions

Open Access This article is licensed under a Artistic Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long equally y'all give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the cloth. If material is not included in the article's Creative Commons licence and your intended employ is not permitted by statutory regulation or exceeds the permitted utilize, you volition demand to obtain permission straight from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/four.0/.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this commodity

Beevers, L., Collet, Fifty., Aitken, G. et al. The influence of climate model uncertainty on fluvial flood hazard estimation. Nat Hazards 104, 2489–2510 (2020). https://doi.org/10.1007/s11069-020-04282-4

Download citation

Received: 07 March 2019
Accepted: 29 August 2020
Published: 09 September 2020
Issue Date: Dec 2020
DOI : https://doi.org/x.1007/s11069-020-04282-four

Keywords

Flood inundation
Climate change
Incertitude quantification
Probabilistic

edwardsintle1976.blogspot.com

Source: https://link.springer.com/article/10.1007/s11069-020-04282-4