Vineyard irrigation management relies on accurate assessment of crop evapotranspiration (ET_c). ET_c is affected by the by type of plant, its physiological properties, and meteorological parameters. Rapid measurement of these factors facilitates quantification of ET_c and enables skilled decision-making for data-driven irrigation. Our main objective was to quantify the performance of different modeling approaches for forecasting seasonal ET_c using meteorological and vegetative data (e.g., leaf area) from five consecutive growing seasons (2013–2017) of Vitis vinifera 'Cabernet Sauvignon' vines. Time series of ET_c was acquired from water balance from vines grown in drainage lysimeters within the vineyard. ET_c forecasts were generated for each season using twelve regression models: six linear and six non-linear multivariate adaptive regression spline (MARS) models. Each regression model constituted a unique combination of variables, some relying on crop coefficient (K_c) and others based on direct ET_c forecasting. The models were trained using data from four growing seasons and compared via measures of coefficient of determination (R²), residual standard deviation, and coefficient of variation. Each model was then tested using ET_c forecasts for a fifth growing season, and compared to the measured ET_c values using correlation, root mean squared error (RMSE), and normalized RMSE. Finally, a mean-seasonal rolling RMSE with a window of 7 days was used to assess the accuracy of the different models. The results show a clear advantage to using non-linear modeling for ET_c forecasting (average RMSE range of 0.81–1.05 vs. 0.64–0.71 mm day⁻¹, respectively). Furthermore, direct forecasting and K_c-based methods yielded similar results, and all models benefited from the incorporation of leaf area data. Similar outcomes were found for the rolling RMSE analysis, with improved model accuracy credited to the inclusion of leaf area, especially early in the season. Our findings confirm that advanced algorithms promote site-specific and location-oriented irrigation management.

葡萄园灌溉管理依赖于作物蒸散量 (ET _c ) 的准确评估。ET _c受植物类型、生理特性和气象参数的影响。快速测量这些因素有利于 ET _{c 的}量化，并使数据驱动灌溉的熟练决策成为可能。我们的主要目标是使用来自Vitis vinifera 'Cabernet Sauvignon' 葡萄藤连续五个生长季节（2013-2017 年）的气象和植物数据（例如，叶面积）来量化预测季节性 ET _c的不同建模方法的性能。ET _{c 的}时间序列来自葡萄园内排水蒸渗仪中生长的葡萄藤的水分平衡。使用十二个回归模型为每个季节生成ET _c预测：六个线性和六个非线性多元自适应回归样条 (MARS) 模型。每个回归模型都构成了一个独特的变量组合，一些依赖于作物系数 (K _c )，而另一些则基于直接的 ET _c预测。这些模型使用来自四个生长季节的数据进行训练，并通过测定系数 (R ² )、残留标准偏差和变异系数进行比较。然后使用 ET _c预测对第五个生长季节进行测试，并与测量的 ET _{c 进行比较}值使用相关性、均方根误差 (RMSE) 和归一化 RMSE。最后，使用窗口为 7 天的平均季节性滚动 RMSE 来评估不同模型的准确性。结果显示了使用非线性建模进行 ET _c预测的明显优势（平均 RMSE 范围分别为 0.81–1.05 和 0.64–0.71 mm day ^-1）。此外，直接预测和 K _c基于的方法产生了类似的结果，所有模型都受益于叶面积数据的结合。滚动 RMSE 分析发现了类似的结果，模型准确性的提高归功于包含叶面积，尤其是在季节早期。我们的研究结果证实，先进的算法可促进特定地点和面向位置的灌溉管理。