Abstract:
This study is carried out to compare three hydrological methods for computing return period of drought events namely : (1) Frequency Analysis (FA), (2) Flow Duration Curve (FDC), and (3) Theory of Run (TR). Runoff data from Ping River at Station P.1 in Chiang Mai, Thailand, is used for which the definitions of drought, the analysis procedures and the application guidelines are compared. Conclusions drawn from the study are : FA method follows the definition of hydrological drought. Daily flow data was used with the minimum average flow of any interval as the main parameter. In case that data series contain some zero flow values, it found that many widely used probability distribution functions cannot fit data properly. Difference periods of water year effected directly the results in term of the return periods. For FDC method, it is also based on the definition of hydrological drought and uses daily flow data in the analysis. FA is applied to identify the return period of flow duration curve. The interested parameter for this method is the average daily flow. Different periods of water year did not have significant effect on the results, i.e. the return periods. For TR, for which drought is defined jointly from hydrological and socioeconomic perspectives, the key parameters are drought severity and drought duration. Data types for the analysis can be daily, monthly or annual flows and these depend on specific problems. The analysis uses concept of joint probability to relate two key parameters (drought severity and drought duration) with return period. It found that gamma distribution is appropriate for fitting the conditional probability of drought severity for specified durations while the probability of drought duration was dictated by the threshold and none of the distribution functions can be used to fit the distribution of drought duration in every cases of threshold. For the application of FA and FDC, the flow threshold must be defined as a constant in order to analyze the relationship between the natural flows and the demands for water uses, while in the case of TR the threshold can be either fixed or varied with time. The return periods determined from these three methods are suitable for applying in different works. FA and FDC are appropriated for determining the capability of channel to divert flow in run-off-river typed project while TR, using monthly flow data, could be used to estimate the capacity of reservoir and reservoir operation.