A long-standing problem in maritime operations and ocean development projects has been the prediction of instantaneous wave energy. Wave measurements collected using an array of freely drifting arrays of Surface Wave Instrument Float with Tracking (SWIFT) buoys are used to test methods for phase-resolved wave prediction in a wide range of observed sea states. Using a linear inverse model in directionally-rich, broadbanded wave fields can improve instantaneous heave predictions by an average of 63% relative to statistical forecasts based on wave spectra. Numerical simulations of a Gaussian sea, seeded with synthetic buoys, were used to supplement observations and characterize the spatiotemporal extent of reconstruction accuracy. Observations and numerical results agree well with theoretical deterministic prediction zones proposed in previous studies and suggest that the phase-resolved forecast horizon is between 1–3 average wave periods for a maximum measurement interval of 10 wave periods for ocean wave fields observed during the experiment. Prediction accuracy is dependent on the geometry and duration of the measurements and is discussed in the context of the nonlinearity and bandwidth of incident wave fields.

海上作业和海洋开发项目中长期存在的问题是瞬时波能的预测。使用带有跟踪浮标（SWIFT）的表面波仪器浮标的自由漂移阵列的阵列收集的波浪测量值被用于测试在大范围观测到的海域中进行相分辨波预测的方法。与基于波谱的统计预测相比，在方向丰富的宽带波场中使用线性逆模型可以使瞬时升沉预测平均提高63％。用合成浮标播种的高斯海域进行了数值模拟，以补充观测资料并描述重建精度的时空范围。观测结果和数值结果与先前研究中提出的理论确定性预测区非常吻合，并且建议在实验期间观测到的海浪场的最大测量间隔为10个波周期，相位分辨的预测范围在1-3个平均波周期之间。预测精度取决于测量的几何形状和持续时间，并在入射波场的非线性和带宽的背景下进行讨论。