Machine-Learning blends of geomorphic descriptors: value and limitations for flood hazard assessment across large floodplains

Abstract. Recent literature shows several examples of simplified approaches that perform flood hazard (FH) assessment and mapping across large geographical areas on the basis of fast-computing geomorphic descriptors. These approaches may consider a single index (univariate) or use a set of indices simultaneously (multivariate). What is the potential and accuracy of multivariate approaches relative to univariate ones? Can we effectively use these methods for extrapolation purposes, i.e. FH assessment outside the region used for setting up the model? Our study addresses these open problems by considering two separate issues: (1) mapping flood-prone areas, and (2) predicting the expected water depth for a given inundation scenario. We blend seven geomorphic descriptors through Decision Tree models trained on target FH maps, referring to a large study area (≈105 km2). We discuss the potential of multivariate approaches relative to the performance of a selected univariate model and on the basis of multiple extrapolation experiments, where models are tested outside their training region. Our results show that multivariate approaches may (a) significantly enhance flood-prone area delineation (overall accuracy: 93 %) relative to univariate ones (overall accuracy: 84 %), (b) provide accurate predictions of expected inundation depths (determination coefficient ≈0.7), and (c) produce encouraging results in extrapolation.

[1]  David G. Tarboton,et al.  Terrain Analysis Using Digital Elevation Models in Hydrology , 2003 .

[2]  Salvatore Manfreda,et al.  Detection of Flood-Prone Areas Using Digital Elevation Models , 2011 .

[3]  R. Lasaponara,et al.  Estimating Flood Characteristics Using Geomorphologic Flood Index with Regards to Rainfall Intensity-Duration-Frequency-Area Curves and CADDIES-2D Model in Three Iranian Basins , 2020, Sustainability.

[4]  Florian Pappenberger,et al.  Operational early warning systems for water-related hazards in Europe , 2012 .

[5]  Salvatore Manfreda,et al.  Flood-prone areas assessment using linear binary classifiers based on flood maps obtained from 1D and 2D hydraulic models , 2015, Natural Hazards.

[6]  Quoc Bao Pham,et al.  Novel hybrid models between bivariate statistics, artificial neural networks and boosting algorithms for flood susceptibility assessment. , 2020, Journal of environmental management.

[7]  David G. Tarboton,et al.  On the extraction of channel networks from digital elevation data , 1991 .

[8]  C. Nataraj,et al.  A Framework for Modeling Flood Depth Using a Hybrid of Hydraulics and Machine Learning , 2020, Scientific Reports.

[9]  P. Bates,et al.  Advances in pan‐European flood hazard mapping , 2014 .

[10]  R. De Risi,et al.  Delineation of flooding risk hotspots based on digital elevation model, calculated and historical flooding extents: the case of Ouagadougou , 2018, Stochastic Environmental Research and Risk Assessment.

[11]  Roland L. Redmond,et al.  An Automated Technique for Delineating and Characterizing Valley-Bottom Settings , 2000 .

[12]  T. Kjeldsen,et al.  A bivariate trend analysis to investigate the effect of increasing urbanisation on flood characteristics , 2017 .

[13]  Efi Foufoula-Georgiou,et al.  Floodplain morphometry extraction from a high-resolution digital elevation model: a simple algorithm for regional analysis studies , 2006, IEEE Geoscience and Remote Sensing Letters.

[14]  B. Pham,et al.  Prediction Success of Machine Learning Methods for Flash Flood Susceptibility Mapping in the Tafresh Watershed, Iran , 2019, Sustainability.

[15]  L. Feyen,et al.  Development and evaluation of a framework for global flood hazard mapping , 2016 .

[16]  Xiaowei Xu,et al.  Multi-criteria decision making approaches for supplier evaluation and selection: A literature review , 2010, Eur. J. Oper. Res..

[17]  Jutta Thielen,et al.  The european flood alert system EFAS - Part 2: statistical skill assessment of probabilistic and deterministic operational forecasts. , 2008 .

[18]  Salvatore Manfreda,et al.  Can the basin morphology alone provide an insight into floodplain delineation , 2008 .

[19]  Salvatore Manfreda,et al.  Investigation on the use of geomorphic approaches for the delineation of flood prone areas , 2014 .

[20]  Pierfranco Costabile,et al.  Comparative analysis of overland flow models using finite volume schemes , 2012 .

[21]  G. Tsakiris,et al.  A hybrid method for flood simulation in small catchments combining hydrodynamic and hydrological techniques , 2016 .

[22]  Francesca Carisi,et al.  Evolution of flood risk over large areas: Quantitative assessment for the Po river , 2015 .

[23]  B. Pham,et al.  A comparative assessment of decision trees algorithms for flash flood susceptibility modeling at Haraz watershed, northern Iran. , 2018, The Science of the total environment.

[24]  Alan K. Zundel,et al.  Review of Automated Floodplain Delineation from Digital Terrain Models , 2001 .

[25]  Salvatore Manfreda,et al.  A GIS tool for cost-effective delineation of flood-prone areas , 2018, Comput. Environ. Urban Syst..

[26]  E. Vivoni,et al.  Investigating a floodplain scaling relation using a hydrogeomorphic delineation method , 2006 .

[27]  S. Kanae,et al.  A high‐accuracy map of global terrain elevations , 2017 .

[28]  A. Castellarin,et al.  Predictive Modeling of Envelope Flood Extents Using Geomorphic and Climatic‐Hydrologic Catchment Characteristics , 2020, Water Resources Research.

[29]  C. Lussana,et al.  Evidence of non‐stationarity in a local climatology of rainfall extremes in northern Italy , 2018 .

[30]  W. Youden,et al.  Index for rating diagnostic tests , 1950, Cancer.

[31]  Philip J. Ward,et al.  Increasing flood exposure in the Netherlands: implications for risk financing , 2014 .

[32]  C. Conoscenti,et al.  A comparison of statistical methods and multi-criteria decision making to map flood hazard susceptibility in Northern Iran. , 2019, The Science of the total environment.

[33]  P. Bates,et al.  Evaluation of 1D and 2D numerical models for predicting river flood inundation , 2002 .

[34]  Salvatore Manfreda,et al.  A digital elevation model based method for a rapid estimation of flood inundation depth , 2019, Journal of Flood Risk Management.

[35]  Salvatore Manfreda,et al.  A web application for hydrogeomorphic flood hazard mapping , 2019, Environ. Model. Softw..

[36]  A. Castellarin,et al.  Changes in seasonality and magnitude of sub-daily rainfall extremes in Emilia-Romagna (Italy) and potential influence on regional rainfall frequency estimation , 2020 .

[37]  L. J. Anthony,et al.  The Cambridge Dictionary of Statistics (2nd ed.) , 2003 .

[38]  J. V. Soares,et al.  HAND, a new terrain descriptor using SRTM-DEM: Mapping terra-firme rainforest environments in Amazonia , 2008 .

[39]  Marcello Sanguineti,et al.  Classifiers for the detection of flood-prone areas using remote sensed elevation data , 2012 .

[40]  Salvatore Manfreda,et al.  Geomorphic classifiers for flood-prone areas delineation for data-scarce environments , 2017 .

[41]  Kwok-wing Chau,et al.  Flood Prediction Using Machine Learning Models: Literature Review , 2018, Water.

[42]  Xiaohong Chen,et al.  Flood hazard risk assessment model based on random forest , 2015 .

[43]  G. Di Baldassarre,et al.  Towards understanding the dynamic behaviour of floodplains as human-water systems , 2013 .

[44]  Hyung-Sup Jung,et al.  Spatial prediction of flood susceptibility using random-forest and boosted-tree models in Seoul metropolitan city, Korea , 2017 .

[45]  Michele Brunetti,et al.  Droughts and extreme events in regional daily Italian precipitation series , 2002 .