Khrystiuk et al. (2026) Decomposition, modeling and forecasting of the time series of discharge of the Desna river using the “bsts” package of the R programming language

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Метеорологія. Гідрологія. Моніторинг довкілля.
• Year: 2026
• Date: 2026-05-20
• Authors: Borys Khrystiuk, Liudmyla Gorbachova
• DOI: 10.15407/meteorology2026.09.062

Research Groups

Not specified

Short Summary

The study evaluates the use of the bsts R package to model and forecast daily discharges of the Desna River at Litky village, identifying a student local linear trend model as the most effective for short-term predictions.

Objective

• To develop and evaluate a qualitative forecasting model for daily river discharges of the Desna River (Litky village) using Bayesian Structural Time Series (bsts).

Study Configuration

• Spatial Scale: Desna River, Litky village.
• Temporal Scale: Daily time series (short-term forecasting).

Methodology and Data

• Models used: bsts package (R), specifically testing components including add.seasonal, add.local.level, add.ar, add.local.linear.trend, add.semilocal.linear.trend, and add.student.local.linear.trend.
• Data sources: Daily discharge time series of the Desna River at Litky village.

Main Results

• The daily discharge time series was decomposed into three components: trend, seasonal, and random fluctuations.
• The optimal model was identified as a bsts model without predictors, utilizing the add.student.local.linear.trend function combined with random fluctuations.
• Model performance was validated using five statistical indicators: residual.sd, prediction.sd, rsquare, relative.gof, and MAPE, as well as accumulated error curves.
• The model is most effective for short-term forecasting during periods of stable low water levels or smooth increases and decreases in discharge.

Contributions

• Demonstrates the application of Bayesian Structural Time Series for hydrological forecasting, establishing a link between statistical trend modeling and physical hydrological concepts such as spring flood decline curves and process inertia.

Funding

Not specified

Citation

@article{Khrystiuk2026Decomposition,
  author = {Khrystiuk, Borys and Gorbachova, Liudmyla},
  title = {Decomposition, modeling and forecasting of the time series of discharge of the Desna river using the “bsts” package of the R programming language},
  journal = {Метеорологія. Гідрологія. Моніторинг довкілля.},
  year = {2026},
  doi = {10.15407/meteorology2026.09.062},
  url = {https://doi.org/10.15407/meteorology2026.09.062}
}