Journal of Agrosystems and Analytics

The experiment explored AET seasonal dynamics of a sugarcane field at the Navsari Agricultural University (NAU) in Navsari using the high-resolution remote sensing to address the deficiency of ground-based micrometeorological instruments, e.g. lysimeters or eddy covariance towers. Surface Energy Balance Algorithm on Land (SEBAL) was run in a Python-GRASS GIS setting that combined Landsat 8 satellite data on land with local weather data. The outputs of SEBAL were checked by comparing the estimates with the cloud-based METRIC-EEFlux algorithm. Phenological measurements using NDVI time-series data monitored the sugarcane growth cycle with the greatest vegetative vigor in June and senescence during the months of November to December. Spatial analysis showed that seasonal AET calculated using SEBAL was between 1459 and 1496 mm. An effective rainfall of 676 mm was taken into consideration to estimate the total irrigation water requirements (IWR) of 783 to 820 mm. Conversely METRIC-EEFlux always gave higher values of AET, with the values falling between 1627 and 1698 mm. The pixel-by-pixel comparison of the two models each day showed a moderate correlation (R2 = 0.63) with a Root Mean square error (RMSE) of 1.40mm/day. SEBAL had a mean error (MBE) of bias of -1.17 mm/day, which shows that it was generally underestimated compared to METRIC-EEFlux. The results notwithstanding these differences in the algorithms, SEBAL is a powerful, well-adjusted tool to measure crop water needs and support sound water management decisions in data deficient areas.

This study investigates the performance of the APSIM and CANEGRO crop models in simulating sugarcane growth and yield under different growing environment and fertilizer application. Field experimental data from the 2023–24 season were employed for model calibration, while validation was conducted using data from the 2024–25 season. During calibration, the CANEGRO model demonstrated limited responsiveness to variations in fertilizer dosage, producing uniform outputs across different nutrient treatments. Model validation results indicate that APSIM outperformed CANEGRO in simulating key agronomic parameters including cane yield, aerial dry biomass, and days to emergence. APSIM achieved a higher coefficient of determination (R² = 0.93), D-index (0.95), and lower RMSE (6.01 t/ha) for cane yield compared to CANEGRO (R² = 0.88, D-index = 0.77, RMSE = 6.29 t/ha).

This study investigates the sensitivity of the APSIM and CANEGRO crop models to key climatic parameters and genetic coefficients in simulating sugarcane growth and yield. Sensitivity analyses to genetic coefficient identified critical genetic parameters influencing crop performance. In the CANEGRO model, MaxPARCE, APFMX, and STKPFMAX were the most influential for yield, biomass, and sucrose content, while LFMAX and SER0 significantly impacted LAI and stalk height. These findings suggest that calibration efforts should prioritize phenological, growth, and yield-related parameters. In contrast, global sensitivity analysis using the APSIM model highlighted RUE4, MSS, and GLN as the most impactful parameters affecting cane yield, commercial cane sugar (CCS), and sucrose accumulation, with RUE and MSS emerging as key contributors to both biomass production and sugar content. Additionally, both models exhibited sensitivity to climatic variables. A rise in temperature resulted in a decline in cane yield, more pronounced in CANEGRO (31.35% reduction at +6 °C) compared to APSIM (15.81%). Increases in solar radiation enhanced cane yield and sucrose dry mass, while reduced radiation had adverse effects. Overall, the findings provide actionable insights for improving model calibration and support cultivar selection and management practices under projected climate change scenarios.