Premium
Water Balance Models in One‐Month‐Ahead Streamflow Forecasting
Author(s) -
Alley William M.
Publication year - 1985
Publication title -
water resources research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.863
H-Index - 217
eISSN - 1944-7973
pISSN - 0043-1397
DOI - 10.1029/wr021i004p00597
Subject(s) - autoregressive model , streamflow , water balance , econometrics , environmental science , series (stratigraphy) , mean squared error , balance (ability) , consensus forecast , climatology , meteorology , statistics , mathematics , geography , engineering , medicine , drainage basin , paleontology , cartography , geotechnical engineering , geology , physical medicine and rehabilitation , biology
Techniques are tested that incorporate information from water balance models in making 1‐month‐ahead streamflow forecasts in New Jersey. The results are compared to those based on simple autoregressive time series models. The relative performance of the models is dependent on the month of the year in question. The water balance models are most useful for forecasts of April and May flows. For the stations in northern New Jersey, the April and May forecasts were made in order of decreasing reliability using the water‐balance‐based approaches, using the historical monthly means, and using simple autoregressive models. The water balance models were useful to a lesser extent for forecasts during the fall months. For the rest of the year the improvements in forecasts over those obtained using the simpler autoregressive models were either very small or the simpler models provided better forecasts. When using the water balance models, monthly corrections for bias are found to improve minimum mean‐square‐error forecasts as well as to improve estimates of the forecast conditional distributions.