Modeling daily evapotranspiration time series based on Non-Linear Autoregressive Exogenous (NARX) method and climate...

Area of Science:

• Hydrology and Climate Science
• Environmental Modeling
• Data-Scarce Region Analysis

Background:

• Flood-prone developing nations often lack essential hydrological data.
• Accurate evapotranspiration modeling is vital for water resource management and disaster preparedness.
• Existing methods may be insufficient in data-deficient environments.

Purpose of the Study:

• To explore innovative methods for modeling daily evapotranspiration time series.
• To investigate causal relationships between available climate variables and evapotranspiration.
• To assess the potential of using air pressure and Niño Sea Surface Temperature (SST) indices for evapotranspiration modeling in data-scarce regions.

Main Methods:

• Utilized Convergent Cross-Mapping (CCM) to test causality between climate variables and evapotranspiration.
• Employed the Non-Linear Autoregressive Exogenous (NARX) method for time series modeling.
• Focused on a flood-prone, data-deficient region in Samar, Philippines.

Main Results:

• Convergent Cross-Mapping (CCM) identified direct causal effects from air pressure and four Niño Sea Surface Temperature (SST) indices on evapotranspiration.
• Rainfall was found to have no direct causal effect on evapotranspiration in this analysis.
• The combination of air pressure, Niño SST indices, and the NARX model effectively modeled daily evapotranspiration.

Conclusions:

• Air pressure and Niño Sea Surface Temperature (SST) indices are significant predictors of daily evapotranspiration.
• The developed modeling approach offers a viable solution for evapotranspiration estimation in data-limited, disaster-prone areas.
• This study advances evapotranspiration modeling by highlighting novel causal relationships and a practical methodology.