Scholarly record
COMPARATIVE ASSESSMENT OF RAINFALL INTENSITY-DURATION-FREQUENCY (IDF) RELATIONSHIPS IN BULGARIA: FROM EMPIRICAL METHODS TO DATA-DRIVEN MODELS FOR URBAN DRAINAGE DESIGN
Abstract
This study presents a comparative analysis of three methodologies for deriving intensity–duration–frequency (IDF) relationships of rainfall in Bulgaria: (i) the method embedded in the current regulation for sewerage system design, (ii) the methodology used for estimating maximum river discharge, and (iii) a recently proposed regionalization approach with site-specific IDF curves. The objective is to evaluate their accuracy and applicability under contemporary climatic conditions, with a focus on urban drainage infrastructure. The regulatory method is based on analytical expressions and a simplified two-zone climatic classification. The approach used for maximum river discharge estimation derives rainfall intensity indirectly through reduction relationships from maximum daily precipitation, combined with a more detailed spatial zoning, making it suitable for larger temporal scales. In contrast, the proposed methodology applies statistical modeling of long-term pluviograph data (1961–2020), regional frequency analysis, and refined spatial delineation into twelve homogeneous regions. The resulting IDF curves are compared against reference relationships derived from observed rainfall data at selected meteorological stations. The evaluation employs relative error, bias, and root mean square error (RMSE) across a wide range of durations and return periods. The results indicate that the regulatory approach smooths extreme rainfall values, weakening the characteristic relationships and underestimating intensity variability. The proposed methodology demonstrates the highest accuracy and minimal deviations. These findings support the need for updating national design practices through data-driven regionalization to ensure climate-resilient urban infrastructure.
Publication details
ReferencesPending
Structured references will appear here after the reference import pass. The count is preserved now so the scholarly record is not incomplete.
