Forecasting future estuarine hypoxia using a wavelet based neural network model

摘要: Abstract Ecosystem based modeling and predictions of hypoxia in estuaries their adjacent coastal areas have become increasingly interest to researchers zone managers. Although progress has been made oxygen dynamics short-term predictions, there is still a lack long-term forecasts that incorporate multiple inputs including climatological effects such as El Nino-Southern Oscillation (ENSO) events. In this study, we first develop hypoxic volume index (HVI) using 26-years ( −1 ) measurements from the main-stem Chesapeake Bay. Then cross-wavelet analysis used identify weight input parameters order build neural network model future volume. The time-forward dynamic uses cross-bay winds along with Oceanic Nino Index (ONI), Susquehanna River flow indexes predict over next several years. Wavelet indicates an anti-phase relationship between southwesterly index, 18-month phase lag index. results yield R -values 0.99, 0.91 for training, validation 2 0.68 illustrating usefulness promise these types models Model could be climatologically baseline comparing actual response nutrient load reductions.