Abstract

The identification and adjustment of precipitation time series data for non-climatic changes in recording bias among rain gauges can be instrumental in controlling uncertainty in hydrologic models. Hydrologic models depend upon the reliability of precipitation and flow monitoring data sets used for calibration and simulation. Consistent precipitation and flow monitoring measurements clearly can be important when attempting to characterize rainfall runoff relationships over time. Hydrologic models require rain gauge networks to represent the spatial distribution of precipitation across a drainage basin and benefit from the normalization of relative rain gauge biases across the network.

Calibration of large urban sewer system models, using a moderately-dense basin-wide rain gauge network and continuous flow monitoring data, is improved by creating continuous homogeneous rainfall records with normalized spatial biases.

Double-mass regression and cumulative residual time series analysis techniques are used to evaluate and adjust historical rain gauge network data to correct for non-homogeneity of individual rainfall records and to normalize spatial bias across the network. Homogeneity of rainfall time series data is evaluated and adjusted by comparison to the rain gauge network mean over a 13-year period of record. Spatial bias across the network, then, is normalized by comparison to continuous calibrated radar rainfall estimates obtained over a 15-month period. Cumulative residual time series analysis techniques also are applied to evaluate the homogeneity of flow monitoring data used in model calibration. The benefits of normalizing the rain gauge network biases to model calibration are illustrated by comparing model results using gauge data with and without bias correction.

The identification and adjustment of precipitation time series data for non-climatic changes in recording bias among rain gauges can be instrumental in controlling uncertainty in hydrologic models. Hydrologic models depend upon the reliability of precipitation and flow monitoring data sets used for calibration and simulation. Consistent precipitation and flow monitoring measurements clearly can be important when attempting to characterize rainfall runoff relationships over time. Hydrologic models require rain gauge networks to represent the spatial distribution of precipitation across a drainage basin and benefit from the normalization of relative rain gauge biases across the network.

Calibration of large urban sewer system models, using a moderately-dense basin-wide rain gauge network and continuous flow monitoring data, is improved by creating continuous homogeneous rainfall records with normalized spatial biases.

Double-mass regression and cumulative residual time series analysis techniques are used to evaluate and adjust historical rain gauge network data to correct for non-homogeneity of individual rainfall records and to normalize spatial bias across the network. Homogeneity of rainfall time series data is evaluated and adjusted by comparison to the rain gauge network mean over a 13-year period of record. Spatial bias across the network, then, is normalized by comparison to continuous calibrated radar rainfall estimates obtained over a 15-month period. Cumulative residual time series analysis techniques also are applied to evaluate the homogeneity of flow monitoring data used in model calibration. The benefits of normalizing the rain gauge network biases to model calibration are illustrated by comparing model results using gauge data with and without bias correction.

This paper is only available in PDF Format:

  View full text PDF