Predicting critical period to characterise over-year and within-year reservoir systems

Surface water reservoirs can be classified as either within-year or over-year based on the duration of their critical period (CP). In general, within-year systems are those which will refill and spill several times in a year, whereas over-year systems have much longer critical periods, usually of the order of years. If the duration of the critical period, and hence the precise mode of behaviour, of a reservoir were to be known a priori, then advantage could be taken of this to select the level of detail required for reservoir analysis. For example, if the reservoir system is purely over-year, i.e. the CP is much longer than 12 months, then only annual streamflow data are required for analysis. On the contrary, systems which exhibit dual within-year and over-year behaviours will require time series data of a finer resolution to capture both the seasonal and annual discrepancies between the demand and inflow. Such a consideration often results in a phenomenal increase in the analysis time over that required for annual data. Finally, if the system is purely within-year, then the analysis effort can be significantly reduced by concentrating on the critical or driest year of the record. In this paper, we examine the properties of the test in current use for distinguishing between within-year and over-year behaviours. In particular we investigate how the parameter of the test is related to the CP, and we argue that knowing the CP is a more complete test. We then develop predictive equations for the CP and we offer suggestions for extending the study.