Assessing the Optimistic Bias in the Natural Inflow Forecasts: A Call for Model Monitoring in Brazil

TL;DR

This paper empirically demonstrates a systematic optimistic bias in Brazil's official hydro inflow forecasts, highlighting the need for improved model monitoring and governance to enhance accuracy and reliability.

Contribution

It provides empirical evidence of forecast bias in Brazil's inflow models and discusses sources and mitigation strategies for improved hydro forecast accuracy.

Findings

Significant optimistic bias identified in 14 years of data

Bias increases with forecast horizon, up to 6.73 GW for 24-step ahead

Results suggest current models have limitations requiring better governance

Abstract

Hydroelectricity accounted for roughly 61.4% of Brazil's total generation in 2024 and addressed most of the intermittency of wind and solar generation. Thus, inflow forecasting plays a critical role in the operation, planning, and market in this country, as well as in any other hydro-dependent power system. These forecasts influence generation schedules, reservoir management, and market pricing, shaping the dynamics of the entire electricity sector. The objective of this paper is to measure and present empirical evidence of a systematic optimistic bias in the official inflow forecast methodology, which is based on the PAR(p)-A model. Additionally, we discuss possible sources of this bias and recommend ways to mitigate it. By analyzing 14 years of historical data from the Brazilian system through rolling-window multistep (out-of-sample) forecasts, results indicate that the official forecast model exhibits statistically significant biases of 1.28, 3.83, 5.39, and 6.73 average GW for 1-, 6-, 12-, and 24-step-ahead forecasts in the Southeast subsystem, and 0.54, 1.66, 2.32, and 3.17 average GW in the Northeast subsystem. These findings uncover the limitations of current inflow forecasting methodologies used in Brazil and call for new governance and monitoring policies.