Related Experiment Video
Updated: Oct 22, 2025

Continuous Hydrologic and Water Quality Monitoring of Vernal Ponds
Published on: November 13, 2017
Comparison of Missing Data Infilling Mechanisms for Recovering a Real-World Single Station Streamflow Observation
Thelma Dede Baddoo1,2, Zhijia Li2, Samuel Nii Odai3
1Binjiang College, Nanjing University of Information Science & Technology, No.333 Xishan Road, Wuxi 214105, China.
Accurately reconstructing missing streamflow data is crucial. This study found that while univariate methods work, multivariate imputation algorithms offer comparable or better accuracy for streamflow data without extensive computational costs.
Area of Science:
- Hydrology
- Data Science
- Environmental Science
Background:
- Reconstructing missing streamflow data is essential for hydrological analysis but presents significant challenges, especially with limited auxiliary data.
- Existing imputation algorithm accuracy assessments for real-world datasets are scarce, hindering reliable data reconstruction.
- Understanding the necessary complexity of missing data reconstruction schemes is vital for accurate streamflow data analysis.
Purpose of the Study:
- To investigate the complexity of missing data reconstruction schemes for real-world single-station streamflow data.
- To ascertain the accuracy of various imputation algorithms for streamflow datasets.
- To evaluate the effectiveness of different imputation methods, including univariate and multivariate approaches.
Main Methods:
- Applied diverse missing data mechanisms, from univariate algorithms to multivariate imputation methods.
- Incorporated time as an explicit variable in the imputation models.
- Assessed imputation performance using Total Error Measurement (TEM) and a novel Localized Error Measurement (LEM).
Main Results:
- Univariate algorithms provide satisfactory results for time series streamflow data, but the best-performing ones are computationally intensive.
- Multivariate imputation algorithms, considering data characteristics and surrounding values, achieve comparable or superior performance to univariate methods, with lower computational demands.
- The developed Localized Error Measurement (LEM) is particularly effective for datasets with localized missing data or large gaps.
Conclusions:
- Proper imputation of missing streamflow values requires a thorough understanding of the specific hydroclimatic dataset.
- Multivariate imputation methods offer an efficient and accurate alternative to computationally intensive univariate methods for streamflow data.
- The Localized Error Measurement (LEM) provides valuable insights into imputation accuracy in challenging data scenarios.
Related Concept Videos
Typical Model Studies
Precipitation Titration: Endpoint Detection Methods
In the Volhard method, a standard excess of AgNO3 is first added to the...
Uniform Depth Channel Flow: Problem Solving
Rapidly Varying Flow
Mechanistic Models: Compartment Models in Individual and Population Analysis
Precipitation Gravimetry
In determining nickel by gravimetric analysis, a precipitant of ethanolic dimethylglyoxime is added to a hot nickel salt solution. This is quickly followed by the dropwise addition of dilute ammonia solution until precipitation occurs. A...

