A statistical tool for a hydrometeorological forecast in the lower La Plata Basin

Extreme discharge events in the La Plata Basin need to be prevented. Simple approaches to the forecast problem such as SARIMA models usually predict average values correctly but fail to anticipate extreme events. As an approach to this problem, we used copula methods to model the distribution of the NINO 3.4 index and river streamflow pair. We used this to build a six-months forecast for streamflow 95% percentile using observed index values. We added this forecast as an exogenous variable in a SARIMAX model to predict discharge. Given that NINO events are usually correlated with extreme discharge events, we expected this model to improve the SARIMA model in predicting extreme events. When comparing both models, we effectively found that SARIMAX model is better than a SARIMA model both for 6- and 12-month discharge forecasts in periods when an El Nino event occurs, while it retains the same performance level when evaluated on all the span of the time series. This model emerges as a lightweight and easily implementable option for decision makers to anticipate extreme events and reduce the negative impacts that they generate.

Automated summary

This study used copula methods to model the distribution of the NINO 3.4 index and river streamflow pair and added the forecast of the streamflow 95% percentile as an exogenous variable in a SARIMAX model. The results showed that the SARIMAX model was more accurate in predicting extreme events compared to the SARIMA model during El Nino events.