.png){kind=icon}
.png){kind=icon}
Return
A Framework to Develop High-Resolution Intensity–Duration–Frequency Curves: Historical Analysis and Scaling Relationships Integrating Gauge and Gridded Rainfall Products
.png){kind=icon}
PRE
AI
Y
B
C
DOI：10.1002/joc.70504.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Advances in satellite rainfall retrieval have made remote sensing estimates increasingly reliable for infrastructure design and planning. Intensity–duration–frequency (IDF) curves remain one of the most widely used statistical methods for estimating design rainfall in water resources engineering. In developing countries, meteorological observatories typically have access to daily data and sometimes lack long archives of hourly rainfall data. The current study provides a framework for developing high-resolution IDF curves with durations ranging from 1 to 72 h at various return periods, integrating gauge and gridded rainfall products (GRPs). The duration-dependent generalised extreme value (d-GEV) and Gumbel distributions were evaluated for modelling annual maximum rainfall series (AMS) at stations with hourly observations, with the Gumbel distribution emerging as the better choice. Four GRPs (GPM, GSMaP, PERSIANN and MSWEP) are evaluated at daily and sub-daily time scales by employing the Gumbel distribution on AMS to derive the IDF curves. In the first stage, return levels are estimated for 24-, 48- and 72-h durations by merging a dense network of daily rainfall data from non-recording rain gauge stations with gridded products using machine learning regression techniques. The scale invariance theory of rainfall is investigated using hourly rainfall data from self-recording rain gauge stations (SRRG) to derive rainfall intensities for sub-daily durations. Scaling behaviour was assessed at each SRRG station using non-central moments (NCMs) and was found to perform well for modelling rainfall extremes. The derived scaling exponents are spatially interpolated using potential covariates to develop IDF curves at 10-km resolution. The framework is implemented for the Indian subcontinent, which shows significant variability in rainfall spatial and temporal patterns. The research outputs are available through a user-friendly open web platform, where users can select a location on an interactive map to view IDF curves and confidence intervals via interactive graphs and tables.
Keywords:
design rainfall
intensity–duration–frequency curves
random forest
remote sensing precipitation products
scale invariance
Journal
IF:
2.8
Papers:
7.9K
Citations:
2.8W
Organization
Cited Papers