Improving estimation of missing values in daily precipitation series by a probability density function‐preserving approach

Abstract

This work presents a novel method for estimating missing values in daily precipitation series. It is aimed at identifying the event time location with good accuracy and reconstructing the correct amount of daily rainfall. In addition, the statistical properties of the time series, i.e. both probability distribution and long-term statistics, are preserved. The completion method is based on a two-step algorithm that uses information from a cluster of neighboring stations. First, wet and dry days are tagged, and subsequently, the full precipitation amount for wet-classified days is estimated by a modified multi-linear regression approach. This method avoids overestimation of the number of wet days and underestimation of intense precipitation events, which are typical side effects of common regression-based approaches.

A Hybrid Method for Interpolating Missing Data in Heterogeneous Spatio-Temporal Datasets

“…The missing records can be calculated by the other records of missing observation stations using Equation (2). To ensure thatt 0 is the unbiased estimate for the missing records, the following relationship should be satisfied:…”

“…Nearly all instrumental space-time data are influenced by missing data [1,2]. Before data analysis, missing data must be well handled.…”

“…One is to exclude periods with missing values from data analysis, and the other is to ignore the missing data based on the tacit assumption that the data represent one continuous series [3,4]. However, these approaches may disregard useful information and bias the analysis results [2]. To overcome these limitations, a number of interpolation methods have been proposed to estimate missing observations in space-time data.…”

Constructing a 1-km Gridded Multi-Scalar Drought Index Dataset (1960 - 2012) in Taiwan Based on the Standardized Precipitation Evapotranspiration Index-SPEI

“…The incomplete time series in both rainfall and temperature data found in both daily and monthly-mean records need to be imputed. The technique described in Simolo et al (2010) (and earlier works: Shepard 1968(and earlier works: Shepard , 1984Willmott et al 1985) is simple but has been proved to be efficient in high-resolution data analysis (e.g., Brugnara et al 2012). This method is employed to perform the imputation of missing data in station time series.…”

“…We chose the top 10 reference stations with the highest products of three weights to calculate their weighted average. Interested readers in the data imputation technique can refer to Simolo et al (2010;also see Weng and Yang 2012 for its recent application in Taiwan) for details.…”