Hydrologic Modeling for Flood Frequency Prediction in Several Data-Limited River Basins in the Balkans

Abstract

Flooding is the costliest natural hazard in Europe, and poses a serious threat to life and property across the globe. Consequently, identifying areas prone to flooding under a variety of storm conditions is an important step in quantifying flood risk, and mitigating and preventing flood-related damages and losses. Flood frequency analysis plays a critical role in the identification of flow magnitudes under different recurrence interval scenarios. In ungauged catchments, flood frequency estimates can only be obtained by using regional regression equations or hydrologic models. In regions of the world such as the Balkans, which have only very limited data and lack well-established regional regression equations, hydrologic models play a pivotal role in providing these flood frequency estimates. This study uses a probabilistic hydrologic model to simulate the rainfallrunoff process in three Balkan countries: Albania, Macedonia, and Serbia. The model domain includes all major river basins contributing surface runoff to these three countries. The key model input is derived from soil, land-use and land-cover, and topography data available for the modeled area. The statistically downscaled output from a fully-coupled Global Circulation Model and Numerical Weather Prediction model is used as the hydrologic model’s precipitation input. This hourly rainfall data set has a resolution of 8.18 km. Simulated flow is used to generate flood frequency estimates for both gauged and ungauged streams within the three countries. Results show that the approach has great promise for accurately predicting flood frequency values in the region of interest.