Spatial prediction of flood susceptible areas using rule based decision tree (DT) and a novel ensemble bivariate and multivariate statistical models in GIS
Introduction
Flood is considered as a severe natural hazard and the coverage of its damages is not measurable (Rozalis et al., 2010). Kron (2002) describes flooding as a result of heavy precipitation and snow melting that makes the rivers overflow form their normal border and temporarily covers the land which was not used to be covered by water. This type of flooding is classified as river flood. While there are two other types of flash flood and coastal flood exist, but river flood can be predicated through proper methods (Jonkman, 2005). In natural hazard management especially in flood management time is one of the most important factors i.e. the employed model should be fast in order to assist the early warning and prevention measures. Many studies have been done in order to measure and classify the flood impacts from various perspectives. Generally, damages can be direct and indirect, or tangible and intangible which all should be considered in flood damage assessment (Merz et al., 2004, Smith and Ward, 1998). Opolot (2013) stated that between 2000 and 2008 almost 99 million people per year were affected by flood alone worldwide. The frequent increase of flood events are mainly due to rapid urbanization and civilization along the rivers, and also cutting the forests (Bronstert, 2003, Christensen and Christensen, 2003). For that reason, susceptible areas to the flood should be detected in order to avoid generating more development in these areas and also to be able to have fast emergency response in various circumstances.
High frequency of the flood occurrence in Malaysia made this disaster as the most important natural hazard causing many deaths, loss of properties and damages to the ecosystem (Pradhan and Youssef, 2011). Since the 1920s many reports have been recorded about the flood occurrences in Malaysia. Department of Irrigation and Drainage (DID) stated that 9% of land area (29,800 km2) in Malaysia is susceptible to flood and also 22% of the population (4.82 million) is affected by this disaster (Pradhan, 2010a). Kelantan is one of the 13 states of Malaysia and is highly affected by the annual monsoon floods during last decades. Recently, heavy monsoons rainfall has triggered floods in Malaysia and especially threatens some states such as Kelantan, Terengganu and Pahang which are located along Malaysia’s east coast (Pradhan and Youssef, 2011). The flood cost nearly million dollars of property and many lives which could have been prevented or mitigated if an early warning system was in place. Human activities such as interference in natural cycle by land use/cover (LULC) changes, unplanned urban expansion near to the bank of the rivers, and uncontrolled construction of buildings can influence the spatial and temporal pattern of hazards. Therefore, an assessment of the basin structure, climate condition, and susceptible areas, assists to prevent the damages which threat the human lives and properties.
Usually, flood management can be done through four stages: prediction, preparation, prevention and damage assessment (Konadu and Fosu, 2009). The efficiency of RS and GIS made the revolution in hydrology and specially flood management which could fulfill all the requirements of each stage. Different types of analyses can be done prior to the flood occurrence, during and after its event. Traditional flood models are increasingly improved or replaced by rule-based and automated methods which are more robust in hazard analyses (Hostache et al., 2013). So floods can be predicted and the flood risk and vulnerable areas can be mapped out. Through susceptibility analysis the areas which have high potential to the flooding can be recognized and therefore; early warning and emergency response can be performed in order to facilitate early preparations and decrease the effects of the disaster (Kia et al., 2012).
Section snippets
Previous studies
In the recent years, many methods have been developed and applied in flood susceptible mapping. Some hydrological models such as WetSpa (Liu and De Smedt, 2004), HYDROTEL (Fortin et al., 2001) and SWAT (Jayakrishnan et al., 2005) are integrated with RS and GIS for the purpose of data collection and spatial analysis. But more robust and automated methods are needed to be used in order to solve the disadvantages of the traditional hydrological methods (Li et al., 2012). Townsend and Walsh (1998)
Study area and data
This study focuses on the part of Kelantan River Basin in North East part of Peninsular Malaysia, covering 923 km2. Kelantan state is one of the 13 states of Malaysia and Kelantan River is the major river in Kelantan state. About 85% of the Kelantan state’s has been covered by the basin.
Due to the geographical location of the study area, it attracts much tourism. Unfortunately monsoon floods affect the area every year. In this study, the flood event of November 2005 was used as flood inventory.
Statistical analysis
The use of flood susceptibility maps for the purpose of LULC planning has increased significantly during the last few decades (Cerra and Prange, 2012). Such mapping assists to recognize and categorize the areas which are threatened by present or future flooding. In this paper, statistical analysis was chosen. For modeling purpose, both DT and ensemble FR and LR methods were used to compare and evaluate their efficiency in flood susceptibility mapping.
Flood susceptibility mapping using integrated FR and LR model
The first results achieved using FR method represents the weights for each classes of each conditioning factor. The FR for 15 factors was calculated from their relationship to the flood events, as shown in Table 1.
Through FR analysis the relationship between flood occurrence and the classes of each conditioning factor was found. As can be seen in the Table 1, the results of DEM analysis showed that the lowest class (2–19) was the most influential class and the highest class (230.1–865) was the
Conclusion
Flood is one of the serious catastrophes that the human society is facing for many decades. Many attempts have been made in order to control and mitigate it during the last decades. Most of such actions and strategies are practiced in developed countries in which large amount of spatial database are available for analysis. Susceptible areas to the flooding should be detected to predict its spatial distribution for future events. This study addressed flood susceptibility mapping using data
References (82)
- et al.
An easy-to-use MATLAB program (MamLand) for the assessment of landslide susceptibility using a Mamdani fuzzy algorithm
Comput. Geosci.
(2012) - et al.
The application of GIS-based logistic regression for landslide susceptibility mapping in the Kakuda-Yahiko Mountains, Central Japan
Geomorphology
(2005) - et al.
Spatial soil zinc content distribution from terrain parameters: a GIS-based decision-tree model in Lebanon
Environ. Pollut.
(2010) - et al.
Combining GIS with fuzzy multicriteria decision-making for landfill siting in a fast-growing urban region
J. Environ. Manage.
(2008) - et al.
Flood estimation at ungauged sites using artificial neural networks
J. Hydrol.
(2006) - et al.
Classifiers for the detection of flood-prone areas using remote sensed elevation data
J. Hydrol.
(2012) Applicability of neuro-fuzzy techniques in predicting ground-water vulnerability: a GIS-based sensitivity analysis
J. Hydrol.
(2005)- et al.
Urban flood hazard zoning in Tucumán Province, Argentina, using GIS and multicriteria decision analysis
Eng. Geol.
(2010) - et al.
Scheduling satellite-based SAR acquisition for sequential assimilation of water level observations into flood modelling
J. Hydrol.
(2013) - et al.
Assimilation of spatially distributed water levels into a shallow-water flood model. Part II: Use of a remote sensing image of Mosel River
J. Hydrol.
(2010)