Improved weighting methods, deterministic and stochastic data-driven models for estimation of missing precipitation records

Teegavarapu, Ramesh S. V.; Chandra-Mouli, Venkatraman

doi:10.1016/j.jhydrol.2005.02.015

Cited by 288 publications

(208 citation statements)

References 18 publications

Supporting

Mentioning

191

Contrasting

Unclassified

Order By: Relevance

“…Therefore, the hourly rainfall data collected from these three stations between 1970 and 2009 were used to estimate the regional average rainfall p for the watershed by using the reciprocal-distance-squared method (Chow et al, 1988). Since this method is simple and can directly reflect the weighting of distance, it has been widely used not only in determining ungauged or regional average rainfalls but also in determining unmeasured or regional average physical quantities in other fields, such as hydrology and earth science (Ashraf et al, 1997;Cheng, 1998;Teegavarapu and Chandramouli, 2005;Chen, 2011). The regional average rainfall estimated by the reciprocal-distance-squared method can be expressed as follows:…”

Section: Regional Average Rainfall In the Chenyulan Stream Watershedmentioning

confidence: 99%

Recent changes in the number of rainfall events related to debris-flow occurrence in the Chenyulan Stream Watershed, Taiwan

Chen

Huang

Jan

et al. 2012

Nat. Hazards Earth Syst. Sci.

View full text Add to dashboard Cite

Abstract. This study analyzed the variability in the number of rainfall events related to debris-flow occurrence in the Chenyulan stream watershed located in central Taiwan. Rainfall data between 1970 and 2009 measured at three meteorological stations nearby/in the watershed were collected and used to determine the corresponding regional average rainfall for the watershed. Data on debris-flow events between 1985 and 2009 were collected and used to study their dependence on regional average rainfall. The maximum 24-h regional rainfall R d was used to analyze the number of rainfall events N r , the number of rainfall events that triggered debris flows N d , and the probability of debris-flows occurrences P . The variation trends in N r , N d and P over recent decades under three rainfall conditions (R d > 20,230, and 580 mm) related to debris-flow occurrence were analyzed. In addition, the influences of the Chi-Chi earthquake on N d and P were presented. The results showed that the rainfall events with R d > 20 mm during the earthquakeaffected period (2000)(2001)(2002)(2003)(2004) strongly responded to the increases in the average number of rainfall events that triggered debris flows and the average probability of debris-flows occurrences. The number of rainfall events with R d > 230 mm (the lower boundary for the rainfall ever triggering debris flow before the Chi-Chi earthquake), and R d > 580 mm (the lower boundary for extreme rainfall ever triggering numerous debris flows) in the Chenyulan stream watershed increased after 2000. The increase in the number of extreme rainfall events with R d > 580 mm augmented the number of rainfall events ever triggering numerous debris flows in the last decade. The increase in both the number of rainfall events that ever triggered debris flows and the probability of debrisflow occurrences was greater in the last decade

show abstract

Section: Regional Average Rainfall In the Chenyulan Stream Watershedmentioning

confidence: 99%

Recent changes in the number of rainfall events related to debris-flow occurrence in the Chenyulan Stream Watershed, Taiwan

Chen

Huang

Jan

et al. 2012

Nat. Hazards Earth Syst. Sci.

View full text Add to dashboard Cite

show abstract

“…The performance of the methods was compared using widely recognized and commonly used error measures, root mean squared error (RMSE) and mean absolute error (MAE), applied to the monthly station data. The RMSE gives relatively high weight to large errors, while MAE gives equal weight to all individual differences (Teegavarapu and Chandramouli 2005).…”

Section: Estimation Of Mean Areal Rainfallmentioning

confidence: 99%

Modelling catchment inflows into Lake Victoria: uncertainties in rainfall–runoff modelling for the Nzoia River

Kizza

Rodhe

et al. 2011

Hydrological Sciences Journal

View full text Add to dashboard Cite

Climate and soil characteristics vary considerably around the Lake Victoria basin resulting in high spatial and temporal variability in catchment inflows. However, data for estimating the inflows are usually sparsely distributed and error-prone. Therefore, modelled estimates of the flows are highly uncertain, which could explain early difficulties in reproducing the lake water balance. The aim of this study was to improve the estimates of catchment flow to Lake Victoria. The WASMOD model was applied to the Nzoia River, one of the major tributaries to Lake Victoria. Uncertainty was assessed within the GLUE framework. During calibration, log-transformation was performed on both simulated and observed flows. The results showed that, despite its simple structure, WASMOD produces acceptable results for the basin. For a Nash-Sutcliffe efficiency (NS) threshold of 0.6, the percentage of observations bracketed by simulations (POBS) was 74%, the average relative interval length (ARIL) was 0.93, and the maximum NS value was 0.865. The residuals were shown to be homoscedastic, normally distributed and nearly independent. When the NS threshold was increased to 0.8, POBS decreased to 54% with an improvement of ARIL to 0.49, highlighting the effect of the subjective choice of likelihood threshold.Key words rainfall-runoff modelling; GLUE; Nash-Sutcliffe efficiency; Nzoia River; Lake Victoria; WASMOD; uncertainty Modélisation des apports de bassin au Lac Victoria: les incertitudes dans la modélisation pluiedébit du Fleuve Nzoia Résumé Les caractéristiques du climat et du sol varient considérablement dans le bassin du Lac Victoria, résultant en une forte variabilité spatiale et temporelle des apports du bassin. Cependant, les données pour l'estimation des entrées sont généralement clairsemées et sujettes à erreur. Par conséquent, les estimations modélisées des débits sont très incertaines, ce qui pourrait expliquer les premières difficultés à reproduire le bilan en eau du lac. Le but de cette étude était d'améliorer les estimations des apports de bassin au Lac Victoria. Le modèle WASMOD a été appliqué au fleuve Nzoia, l'un des principaux affluents du Lac Victoria. L'incertitude a été évaluée dans le cadre de GLUE. Pendant le calage, une transformation logarithmique a été réalisée sur les débits à la fois simulés et observés. Les résultats ont montré que, malgré sa structure simple, WASMOD produit des résultats acceptables pour le bassin. Pour un seuil de 0.6 sur l'efficacité de Nash-Sutcliffe (NS), le pourcentage d'observations compris entre des simulations (POBS) a été de 74%, la longueur de l'intervalle relatif moyen (ARIL) était de 0.93, et la valeur maximale du NS a été de 0.865. On a montré que les résidus étaient homoscédastiques, normalement distribués et presque indépendants. Lorsque le seuil sur NS a été augmenté à 0.8, POBS a diminué à 54% avec une amélioration de l'ARIL à 0.49, soulignant l'effet du choix subjectif de seuil de vraisemblance.

show abstract

“…elevation) were proposed by Daly et al (2002). The coefficient of correlation weighting method (CCWM) proposed by Teegavarapu and Chandramouli (2005) and adopted in a few recent studies (Westerberg et al 2010) provides a method using gauge-to-gauge correlations. Optimization methods for objective selection of raingauges, optimal cluster identification and improvised traditional distance and quadrant-based methods were recently reported by Teegavarapu (2012).…”

Section: Introductionmentioning

confidence: 99%

“…Variants of the inverse distance weighting method (IDWM) or reciprocal distance weighting method have been provided by Teegavarapu and Chandramouli (2005) and Lu and Wong (2008). Teegavarapu (2009) used an association rule mining (ARM) approach to improve estimates of missing precipitation data.…”

Section: Introductionmentioning

confidence: 99%

“…The study confirmed the one main disadvantage of multiple regression was negative estimates. Applications of artificial neural networks (ANNs) for estimation of missing precipitation data were documented by Teegavarapu and Chandramouli (2005) and Teegavarapu (2007). Recent work of Teegavarapu et al (2009) has focused on the development of optimal functional forms using genetic algorithms and mathematical operators for estimating missing precipitation data.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Missing precipitation data estimation using optimal proximity metric-based imputation, nearest-neighbour classification and cluster-based interpolation methods

Teegavarapu

2014

Hydrological Sciences Journal

View full text Add to dashboard Cite

To cite this article: Ramesh S. V. Teegavarapu (2014) Missing precipitation data estimation using optimal proximity metric-based imputation, nearest-neighbour classification and cluster-based interpolation methods, Hydrological Sciences Journal, 59:11, 200959:11, -202659:11, , DOI: 10.108059:11, /02626667.2013 To link to this article: https://doi.org/10.1080/02626667.2013.862334 Abstract New optimal proximity-based imputation, K-nearest neighbour (K-NN) classification and K-means clustering methods are proposed and developed for estimation of missing daily precipitation records. Mathematical programming formulations are developed to optimize the weighting, classification and clustering schemes used in these methods. Ten different binary and real-valued distance metrics are used as proximity measures. Two climatic regions, Kentucky and Florida, (temperate and tropical) in the USA, with different gauge density and network structure, are used as case studies to evaluate the new methods. A comprehensive exercise is undertaken to compare the performances of the new methods with those of several deterministic and stochastic spatial interpolation methods. The results from these comparisons indicate that the proposed methods performed better than existing methods. Use of optimal proximity metrics as weights, spatial clustering of observation sites and classification of precipitation data resulted in improvement of missing data estimates.

show abstract

Improved weighting methods, deterministic and stochastic data-driven models for estimation of missing precipitation records

Cited by 288 publications

References 18 publications

Recent changes in the number of rainfall events related to debris-flow occurrence in the Chenyulan Stream Watershed, Taiwan

Recent changes in the number of rainfall events related to debris-flow occurrence in the Chenyulan Stream Watershed, Taiwan

Modelling catchment inflows into Lake Victoria: uncertainties in rainfall–runoff modelling for the Nzoia River

Missing precipitation data estimation using optimal proximity metric-based imputation, nearest-neighbour classification and cluster-based interpolation methods

Contact Info

Product

Resources

About