the study of relationship between intensity, duration, frequency and involvement level with drought in the northwest of iran

1. introduction droughts are long-term phenomena that can affect large areas and cause severe downturn in economic and social activities. recognizing this phenomenon, its features and properties, knowledge of facilities, capabilities and talents of areas and providing short and long term solutions leads to the effort of the surface water resource managers to prepare for drought mitigation. the regional distribution of drought is its most notable feature. accordingly, regional analysis of drought conditions compared to its station survey results in a much more complete understanding of this phenomenon. therefore, awareness of the severity, duration and extent of the area affected by the drought and the possibility of its occurrence in the certain return period can be useful in the management of water resources and the condition in which these vital resources are threatened (water, soil, forest and pasture). this can help managers in making appropriate decisions to minimize the damage caused by years of drought. determining the characteristics of intensity, duration, frequency and level of involvement with the drought and the relationship between them can be a great help for authorities and researchers towards appropriate management of surface water resources, soil, agriculture, environmental, urban and rural planning and the tourist. according to the importance of drought in the north west of iran in recent years, we have decided to identify and evaluate drought features to deal with this phenomenon correctly. 2. the study area the study area is located in western north between 440220to 480222 eastern longitude and 355805to 394616 northern latitude this area surrounds eastern and western azerbaijan and ardebil provinces. the researched area included two areas 0f h1 and h2 namely some parts of caspian sea and orumie lake from hydrological view. 3. material and methods this study was conducted to investigate the meteorological situation of drought in the north west of iran by selecting monthly rainfall data from the studied stations among synoptic stations with 22 years of data in terms of geographical location and climatic conditions with proper distribution in the area. the required information was prepared by the meteorology and water management organization. by collecting data from 20 stations for 22 years (october 1988 to september 2009), a data file was prepared in the required format by the excel software. to identify, analyse and monitor the drought in the above stations, the standardized precipitation index, spi was used. mckee et al (1993) determined a threshold for drought in a time scale. drought occurs when spi consistently becomes negative or highly reaches to -1. the drought ends when spi becomes positive. therefore, a drought has a period of which the beginning and ending and the intensity are determined based on the values obtained in the continuous ongoing months. in order to calculate the spi, the mean, standard deviation and skewness of monthly precipitation data is calculated; then, the logarithm of precipitation data is taken and the data logarithmic mean is calculated. using the defined relations, scale factor and the form α and β are calculated. then, data is fitted to gamma function; finally, drought index value is determined by abramowitz-stegun approximation. therefore, the spi is obtained from displacement t by various formulas according to magnitude of a displaced gamma value. the numbers obtained to determine the 3, 6, 12, 18, 24 and 48-month droughts are also used. 4. results and disscution the results show that the overall frequency of 1, 2 and 4 month droughts have the most iteration in all stations and the frequency of, 3, 5, 6, 7, 8 and 9 month droughts have the least iteration. the regression equations are used to estimate the relationship between frequency and duration of drought in the stations of north west. equations obtained by the independent variable, duration, show that the trend was reversed between the two stations and does not merely follow a one-order linear equation. for a sudden increase in the frequency of droughts in 4 and 5 months changes the general trend of decreasing and removes the linear mode. on the other hand, decreased frequency is not proportional to increased duration of drought and there are various slopes. by fitting frequency and duration data to non-linear equations, as a result, r2 can be considered as the suitable equation criteria for determining the severity of the relation. r2, that is the ratio of the square of the predicted variance to the variance of the observed values , is higher. in other words, the coefficient shows how much of the total variance of the observed variable can be explained by the variance of the predicted values by the regression model. creating a 264*20 matrix by spss, wet and drought coefficients are grouped. with this method, the stations are divided into 4 clusters (separate groups) in terms of wet-drought features. based on spi values, stations included piranshahr, sardasht, saghez, mahabad, orumieh, maragheh, tekab, zarina obato and zanjan in the first group, astara and pars abad (group ii), ardebil, ahar, khalkhal, mianeh, tabriz, sarab (group iii) and julfa, khoy and maku in group iv. hence, wet and drought conditions in terms of both spatial and temporal stations are almost identical in stations of a group and differ between groups. as noted in the theoretical discussion of the drought characteristics, onset, termination, intensity and level of involvement is very important in terms spatial and temporal conditions. in this regard, an 8 month period (february to september 2007) was selected and wet-drought classes were zoned on the studied region in the form of 8 maps using gis. reviewing the drought monitoring occurred in september 2007, it is clear that a large part of northern west in september was involved with the drought in the defined ranges, while droughts in september covered total area in all 4 periods. the above analysis has been performed for other months of 2007. 5. conclusion drought frequency has mostly occurred in the weak and moderate intensities during the considered time cycle. the frequency-duration relation indicates a reverse correlation between two characteristics of droughts. the results obtained from analysing droughts occurred in the north west of iran indicate that frequency of 1-, 2- and 4-month droughts was most-iterated frequency in all stations. therefore, it is not clear that all short-time droughts are more than long-time droughts. for example, the frequency of 3-month drought is lower than that of 4-month drought. correlation of spi values at 20 studied stations indicates synchronized drought in adjacent stations; the rate decreases with increasing distance. given that the equation is of 6 orders in most of the times, the curve is sinusoidal and nonlinear. on the other hand, decreased frequency is not proportional to increased duration and different slopes can be observed. therefore, the duration-frequency relation cannot be linear. in addition, grouping the stations into four separate groups suggests an extra-group adifference between stations of a group and little or no difference between inter-groups. hence, wet and drought spatial and temporal conditions are almost similar in stations of a group and differ between groups. monitoring drought occurred in september 2007, finally, it is clear that a large part of northern west was involved with spi drought in the september within the defined range; according to maps, it is not clear that spatial behaviour of drought was the same in all stations during drought.