Repo for R code on forecast skill worked example in 2nd edition drought book.
The forecast dataset ("forecast_data.csv"): hindcasts from the Ensemble Streamflow Prediction (ESP) forecast method for the Lambourn catchment at Shaw gauging station in southern England (catchment area of 234.1 km2). The dataset consists of monthly-averaged river discharge observations in the second column and corresponding ESP forecasts initialised on 1st January each year for a one month lead time for 51 ensemble members, in columns 3 to 53. The dataset contains 50 observation-forecast pairs across the 1965 to 2014 hindcast periods as shown in column 1. The units for both observations and forecasts are m3/s. More detail on how these data were generated can be found in Harrigan et al. (2018).
Data source: Harrigan, S., Prudhomme, C., Parry, S., Smith, K. and Tanguy, M.: Benchmarking ensemble streamflow prediction skill in the UK, Hydrology and Earth System Sciences, 22(3), 2023–2039, doi:https://doi.org/10.5194/hess-22-2023-2018, 2018.