Floods are a type of natural calamity that can harm infrastructure, socioeconomics, and human lives. To offer citizens a sustainable flood risk management system, flood forecasting is crucial. This paper suggests a straightforward machine learning (ML) method that consists of two or more generic algorithms that work in conjunction to answer problems that they were not intended to. Since the majority of machine learning algorithms are tailored for a specific dataset or task, merging different ML algorithms can significantly enhance the end result by assisting in either tuning one another, generalisation, or adaptation to new tasks. The purpose of this paper is to provide an understanding and comprehensive review of machine learning-based hybrid models used in long-term and short-term flood forecasting. It entails researching machine learning-based hybrid models used for flood forecasting and conducting a comparative assessment of the models' parameters, pre-processing methods, and performance measurements. According to this review, machine learning-based hybrid models have been widely used for short-term and long-term flood forecasting. As predictors, various parameters or flood variables have been used. The hybridization of the model has been found to improve forecast performance. The findings of this study will benefit future researchers by providing information on current progress in the use of machine learning-based hybrid models in short-term and long-term flood forecasting.

Flood Forecasting, Hybrid Models, Machine Learning Models, Long-term and Short-term Flood Forecasting.