Abstract:
Sustainable irrigation development is essential for ensuring food security in Bangladesh, especially with the challenges posed by climate change. This research explores the possibility of obtaining reliable high-resolution estimates of evapotranspiration (ET) using remote sensing technology in the northwest region of Bangladesh. To achieve this, a data fusion technique is used to retrieve high-resolution Land Surface Temperature (LST) data, which is then compared with air temperature data from the Bangladesh Meteorological Department (BMD) stations. The Surface Energy Balance Algorithm for Land (SEBAL) model is employed to calculate ET from the fused LST data, and its performance is compared with observed data from lysimeter located at Bangladesh Wheat and Maize Research Institute (BWMRI), Dinajpur. Additionally, spatiotemporal mapping of ET is presented to visualize variations in ET across the same area where the lysimeter is located. The study finds a stronger correlation between maximum air temperature (Tmax) and the fused LST compared to minimum and average air temperature. For Dinajpur, the coefficient of determination (R²) values for the fused LST product are 0.85, 0.74, and 0.81 when compared to Tmax, Tmin, and Tavg, respectively. And for Bogra, the coefficient of determination (R²) values for the fused LST product are 0.75, 0.69, and 0.74 when compared to Tmax, Tmin, and Tavg, respectively. Seasonally, the correlation between Tmax and LST is stronger during the dry season, while the correlation between average air temperature and LST is better during the summer to monsoon season. However, the study notes that the Rangpur station exhibits poor correlations (R² = 0.37) between Tmax and fused LST. Regarding ET estimation for wheat fields and Boro rice, the SEBAL model tends to underestimate ET compared to ground-based measurements, but it still demonstrates reasonable trends. Mean absolute error (MAE) was found 0.16 mm/day between the model ET and observed ET. The research concludes that remote sensing technology, particularly data fusion and the SEBAL model, holds promise for obtaining accurate high-resolution estimates of ET in the northwest region of Bangladesh. However, careful consideration of station-specific variations and temperature variables is necessary for reliable results. The findings contribute to the understanding of evapotranspiration dynamics in the region and offer insights for effective irrigation scheduling and water management.