Analysis of groundwater-level fluctuation and linear regression modeling for prediction of initial groundwater level during irrigation of rice paddies in the Nasunogahara alluvial fan, central Japan

• Published in: Environmental earth sciences Vol. 82; no. 20; p. 473
• Main Authors: Tsuchihara, Takeo, Yoshimoto, Shuhei, Shirahata, Katsushi, Nakazato, Hiroomi, Ishida, Satoshi
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.10.2023; Springer Nature B.V
• Subjects: Air temperature; Alluvial fans; Antecedent precipitation; Aquifers; Biogeosciences; Cereal crops; Climate change; Drought; Earth and Environmental Science; Earth Sciences; Environmental risk; Environmental Science and Engineering; Geochemistry; Geology; Groundwater; Groundwater irrigation; Groundwater levels; Hydrology/Water Resources; Irrigation; Land use; Mathematical models; Modelling; Nitrates; Observation wells; Original Article; Periodicity; Precipitation; Predictions; Regression analysis; Regression models; Rice; Rice fields; Risk assessment; Seasonal variations; Seasons; Statistical analysis; Terrestrial Pollution; Trends; Water shortages; Water supply; Water use; Weather forecasting; Japan; Regression modeling; Paddy rice field; Groundwater; Drought; Alluvial fan; Trend analysis
• ISSN: 1866-6280; 1866-6299
• DOI: 10.1007/s12665-023-11174-w

This study analyzed long-term fluctuations of groundwater levels in six shallow observation wells in the Nasunogahara alluvial fan, Japan’s second largest source of agricultural irrigation groundwater, and presented a simple method for predicting groundwater levels in April prior to the annual planting of paddy rice. The 22-year time-series of groundwater levels (1998–2019) clearly showed seasonal periodicity, with higher levels in summer than in winter. In particular, groundwater levels were lowest in April when groundwater demand was greatest. Groundwater levels in two wells at the beginning of the April irrigation period showed long-term declining trends that can be attributed more to changes in land use than to changes in precipitation or air temperature. A simple linear regression of mean groundwater level in April to antecedent precipitation provided reasonable predictions of April groundwater levels, which were significantly influenced by precipitation in the preceding 3–5 months. Further modeling after subtraction of long-term seasonal trends (detrending) improved these estimates. The performance of the linear regression model for prediction of April groundwater levels is comparable to that of the statistical benchmark model. Using long-term monthly or seasonal weather forecasts, the modeling presented here can be applied to inform appropriate changes of water use practices, such as decreasing groundwater extraction by implementing rotational water supply, changing rice-cropping seasons, or targeting deeper aquifers. The identification of the critical period of antecedent precipitation that affected April groundwater levels in the Nasunogahara alluvial fan is also important for understanding appropriate precipitation periods to be targeted in modeling for future drought risk assessments under climate change.