Investigation of patterns of precipitation temporal distribution in semi-arid regions (Case study: Sanganeh research station, Korasan Razavi province)

Comprehensive management of water resources nowadays demands accurate and complete meteorological and hydrological data. Simulation of runoff hydrograph for hydraulic structures design often requires rainfall characteristics such as the amount, duration, and temporal rainfall distribution. The design storm is one of the main factors in flood estimation in watersheds. Considering the temporal variations in heavy storms, precise recognition of the rainfall temporal pattern is very important. Temporal rainfall distribution is one of the significant features of rainfall affecting the quantity and quality of runoff and erosion, which has unfortunately attracted less attention. Temporal rainfall distribution means the amount of rainfall volume per temporal periods of precipitation. In fact, the temporal rainfall distribution pattern indicates the changes in rainfall intensity. The effect of rainfall characteristics, especially temporal rainfall distribution and its intensity, on simulation of runoff events is more understandable. A huge part of errors in the simulation of rainfall-runoff models is due to the uncertainties regarding regions’ rainfall distribution patterns. Therefore, provision of rainfall distribution pattern for each region by using rainfall data is very important since the pattern varies in regions with different climates. In the studies in Iran, with regard to the variety of mechanisms causing rainfall events (altitude, proximity to the sea, and general circulation systems), it is impossible to consider a predominant distribution pattern for all parts of Iran. In this regard, presented 6 and 24-hour rainfall patterns of American Soil Conservation Society have been used. The SCS patterns which are broadly used in Iran and other countries in order to simulate runoff and designate hydraulic structures.
In this research temporal rainfall distribution has been evaluated in the Sanganeh research watershed. The Sanganeh watershed is located in the north-east of Khorasan Razavi province (6015' E, 3641' N), with the area of 50 ha. According to the collected data from the climatologic station and Domartan coefficient of 10.2, the climate of the watershed is classified as semi-arid. The area receives less than 180 mm annual precipitation. In this study, for determination of temporal rainfall distribution Pilgrim method has been used. In this method, in every time duration based on this fact that in which quarter the maximum rain has happened, the first, second, third and fourth rainstorms were defined and temporal rainfall distribution pattern was evaluated by use of the data obtained from rain gauge station. In this study, by investigating all observed storms in the Sanganeh research station in the period of 2005-2013, all individual storms have been derived in the different durations of 9 hr. Then, the Pilgrim method was used based on storms separation in four quarters. By averaging ranks in each quartile, the index rank the final pattern for each time-duration was specified. To statistically evaluate the obtained patterns with durations of interest, Chi square test was used after establishing a contingency table and estimating the rainfall percentage for each quartile and ranking. By averaging ranks in each quartile, index rank of the final pattern of each time-duration was determined.
The obtained results showed that most of the storms which occurred in Sanganeh region were in the first quadrate. So the assumption that the highest amount of precipitation occurs in the middle of time duration is doubtable. Also, with increasing of the storms duration, the number of the storms decreased in the first quadrate and increased in the second to fourth quadrates. Moreover, 61% of erosive storms had the duration of fewer than 3 hours and most of them were in second and third quadrate. Chi square test was used for statistical evaluation of the obtained patterns with distinct durations. The results showed that in the duration of 9 hours, there was a significant relationship between the rank and the quadrate. These patterns had high ability to use while in the others durations, the patterns were not reliable. In addition, comparison of the obtained logical pattern with SCS type patterns showed that only rainfall temporal pattern of the second quadrate rainfalls had a high correlation coefficient with IA type. Thus, as the most of the storms in the Sanganeh watershed were in the first quadrate and had the duration of 1 to 3 hours, using SCS type pattern in this region should be precautionary.