The results of flood frequency analysis are used for many engineering purposes, such as design of flood control structures or delineation of flood plains. Although the flood flow rates are the most common subject of flood frequency analysis, variables like flood elevations or flood volumes can also be of interest for a flood related project and they can also be the subject of frequency analysis.

Statistical methods for frequency analysis are the most straightforward if two basic assumptions are satisfied. First, it is assumed that hydrologic data on extremes represent independent data, i.e. data which is not correlated with adjacent observations. Second assumption is that data are identically distributed, which means that all data are coming from the same population and have the same statistical properties (also known as the assumption on homogeneity). Translated to hydrologic variables, this means that all data reflect the same type of hydrologic processes. With floods, this assumption may be violated if the observed floods were generated in different ways (as a consequence of severe storms or of snow melt). Another aspect of the homogeneity is time independence of flood regime; this assumption may also be violated by changes due to natural factors or human activities. If these basic assumptions about hydrologic data are not fulfilled, more complex statistical methods have to be applied in order to take into account correlation between observed data and their change over time (such as time series analysis).