Comparative Analysis for Estimating Missing Precipitation in Virginia’s Coastal and Mountainous Regions
College
College of Engineering & Technology (Batten)
Department
Civil and Environmental Engineering
Graduate Level
Doctoral
Presentation Type
Poster Presentation
Abstract
Accurate precipitation estimation is essential for hydrological modelling, flood forecasting, and water resource management, particularly in regions where observational data are sparse or incomplete. Despite their critical role, existing studies have not systematically compared statistical methods for estimating missing precipitation, particularly in Virginia’s coastal and mountainous regions. Missing data caused by gauge malfunctions, sparse observation networks, and extreme weather events, can compromise flood predictions, water allocation, and climate resilience efforts. This issue is further exacerbated by the limited spatial and temporal coverage of ground-based observations, as many stations have been established only recently. This study systematically evaluates statistical interpolation methods for estimating missing precipitation data in Virginia, comparing their effectiveness across diverse coastal and mountainous landscapes. Using monthly precipitation records from five stations in the southeastern coastal Hampton Roads region and five in the southwestern mountainous Blue Ridge area, we assess the performance of five interpolation techniques: Normal Ratio Weighted Average (NRWA), Arithmetic Average (AA), Inverse Square-Distance Weighted Average (ISDWA), Reciprocal-Distance Weighting (RDW), and Single Best Estimator (SBE). Complete records serve as reference data to validate the methods using multiple accuracy metrics, including error-based measures, correlation indicators, binary rain/no-rain classification metrics, and relative error evaluations. Additionally, this study examines the seasonal performance of these statistical approaches across both terrains. The results reveal significant regional variations in method effectiveness, providing insights into their applicability in different environmental settings. These findings enhance hydrological modelling and water resource planning by improving precipitation estimates in data-scarce regions. More broadly, this research supports flood risk mitigation and optimized water management strategies in similarly complex environments.
Keywords
Precipitation estimation, Statistical interpolation, Hydrological modelling, Hampton Roads, Blue-Ridge Mountains
Comparative Analysis for Estimating Missing Precipitation in Virginia’s Coastal and Mountainous Regions
Accurate precipitation estimation is essential for hydrological modelling, flood forecasting, and water resource management, particularly in regions where observational data are sparse or incomplete. Despite their critical role, existing studies have not systematically compared statistical methods for estimating missing precipitation, particularly in Virginia’s coastal and mountainous regions. Missing data caused by gauge malfunctions, sparse observation networks, and extreme weather events, can compromise flood predictions, water allocation, and climate resilience efforts. This issue is further exacerbated by the limited spatial and temporal coverage of ground-based observations, as many stations have been established only recently. This study systematically evaluates statistical interpolation methods for estimating missing precipitation data in Virginia, comparing their effectiveness across diverse coastal and mountainous landscapes. Using monthly precipitation records from five stations in the southeastern coastal Hampton Roads region and five in the southwestern mountainous Blue Ridge area, we assess the performance of five interpolation techniques: Normal Ratio Weighted Average (NRWA), Arithmetic Average (AA), Inverse Square-Distance Weighted Average (ISDWA), Reciprocal-Distance Weighting (RDW), and Single Best Estimator (SBE). Complete records serve as reference data to validate the methods using multiple accuracy metrics, including error-based measures, correlation indicators, binary rain/no-rain classification metrics, and relative error evaluations. Additionally, this study examines the seasonal performance of these statistical approaches across both terrains. The results reveal significant regional variations in method effectiveness, providing insights into their applicability in different environmental settings. These findings enhance hydrological modelling and water resource planning by improving precipitation estimates in data-scarce regions. More broadly, this research supports flood risk mitigation and optimized water management strategies in similarly complex environments.
Comments
Authors' Information:
https://www.linkedin.com/in/imiya-chathuranika-38096514b
https://www.odu.edu/directory/dalya-ismael