Application of Hybrid Prediction Methods in Spatial Assessment of Inland Excess Water Hazard

• Published in: ISPRS international journal of geo-information 2020-04, Vol.9 (4), p.268
• Main Authors: Laborczi, Annamária, Bozán, Csaba, Körösparti, János, Szatmári, Gábor, Kajári, Balázs, Túri, Norbert, Kerezsi, György, Pásztor, László
• Format: Article
• Language: English
• Subjects: geostatistics; hybrid spatial prediction model; inundation; land degradation; machine learning; risk mapping
• ISSN: 2220-9964; 2220-9964
• DOI: 10.3390/ijgi9040268

Inland excess water is temporary water inundation that occurs in flat-lands due to both precipitation and groundwater emerging on the surface as substantial sources. Inland excess water is an interrelated natural and human induced land degradation phenomenon, which causes several problems in the flat-land regions of Hungary covering nearly half of the country. Identification of areas with high risk requires spatial modelling, that is mapping of the specific natural hazard. Various external environmental factors determine the behavior of the occurrence, frequency of inland excess water. Spatial auxiliary information representing inland excess water forming environmental factors were taken into account to support the spatial inference of the locally experienced inland excess water frequency observations. Two hybrid spatial prediction approaches were tested to construct reliable maps, namely Regression Kriging (RK) and Random Forest with Ordinary Kriging (RFK) using spatially exhaustive auxiliary data on soil, geology, topography, land use, and climate. Comparing the results of the two approaches, we did not find significant differences in their accuracy. Although both methods are appropriate for predicting inland excess water hazard, we suggest the usage of RFK, since (i) it is more suitable for revealing non-linear and more complex relations than RK, (ii) it requires less presupposition on and preprocessing of the applied data, (iii) and keeps the range of the reference data, while RK tends more heavily to smooth the estimations, while (iv) it provides a variable rank, providing explicit information on the importance of the used predictors.