LAUSR

Inflow forecasts play an integral role in the management and operations of hydropower reservoirs. In Scotland, the horizon of inflow forecasts is limited in range to approximately 2 weeks ahead. Additional forecast information in the sub‐seasonal to seasonal (S2S) range would allow operators to take proactive action to mitigate weather‐related risks, thereby improving water management and increasing revenue. The aim of this study is to develop methods of deriving skilful S2S probabilistic inflow forecasts for hydropower reservoirs in Scotland, without the application of a hydrological model. We forecast inflow for a case study reservoir using a linear regression model, trained on historical S2S precipitation predictions and observed inflow rates. Ensemble inflow forecasts generated from the regression model are post‐processed using Ensemble Model Output Statistics, to create calibrated S2S probabilistic forecasts. We evaluate forecast skill for 11 different horizons, using inflow observations. Probabilistic forecasts of weekly average inflow rates hold fair skill relative to climatology up to 6 weeks ahead (fCRPSS = 0.01). Forecasts of 28‐day average inflow rates hold good skill (fCRPSS = 0.19). The S2S probabilistic inflow forecasts are most skilful during winter, when there is greatest risk of reservoirs spilling. Forecasts struggle to predict high summer inflows even at short lead times. The potential for the S2S probabilistic inflow forecasts to improve water management and deliver increased economic value is explored using a stylized cost model. While applied to hydropower forecasting, the results and methods presented here are relevant to broader fields of water management and S2S forecasting applications.