1 استادیار پژوهشکده حفاظت خاک و آبخیزداری
2 دانشیار گروه جغرافیای طبیعی، دانشگاه تهران
3 استاد گروه جغرافیای طبیعی، دانشگاه تهران
4 استادیار گروه جغرافیای طبیعی، دانشگاه تهران
عنوان مقاله [English]
Regional weather and climate around the globe are strongly influenced by large-scale atmospheric circulation patterns (McKendry, 1994). The classification of atmospheric circulation patterns, as an attempt to reduce the dimensionality of the analysis by isolating a relatively small number of representative circulation types (hereafter CTs), has long been used in synoptic climatology to assess the influence of CTs on climatic variables, including precipitation (Yarnal, 1993; Huth, 1996). Many authors have studied the possible linkages between atmospheric circulation patterns and surface variables, such as precipitation and temperature (Romero et al., 1999; Santos et al., 2005; Kilsby et al., 1998; Wibig, 1999; Chen and Hellstr?m, 1999; Xoplaki et al., 2000; Kidson, 2000). Furthermore, several studies found relationships between CTs and natural hazards, including droughts (Duckstein et al., 1993; Bogardi et al., 1994; Pesti et al., 1996; Pongracz et al., 2003). Therefore, the present study aims at identifying daily 500 hPa CTs and their relative impacts on occurrence of precipitation events in Iran during the winter months.
Materials and Methods
For CT classification, daily means of the 500 hPa geopotential height, defined over a 2.5° latitude × longitude grid and covering an area from the Mediterranean basin to the Middle East (10°E–80°E, 20°N–60°N), were retrieved from the NCEP/NCAR reanalysis (Kalnay et al., 1996; Kistler et al., 2001). The latest version of APHRODITE grided daily precipitation dataset (APHRO_V0902), which is based on 337 Iranian meteorological stations, having record lengths of at least 5 years is also used for identifying the relationship between CTs and precipitation occurrences in Iran. These stations are composed of 154 World Meteorological Organization (WMO) stations and of 183 non-WMO ones, spread throughout the country and covering different temporal periods, with the longest starting in the 1960s (Yatagai et al., 2008).
The S-mode PCA was applied to the correlation matrix of the daily 500 hPa geopotential height and the first six principal components were retained for Varimax rotation after applying the screen test and North’s rule-of-thumb. The spatial variation patterns of the rotated loadings, in their positive and negative phases, i.e. on a location of the dominant centers-of-action, was considered as potential groups for circulation pattern classification in a non-iterative K-means clustering (Esteban et al., 2005, 2006). The centroids are computed by averaging all days that fulfill the “extreme score criterion” for a certain pattern and phase: days with high scores for a certain component (values higher than +2 for positive phase or lower than –2 for negative phase), but with low scores in the remaining components (between +2 and –2) were selected and considered as seeds for the centroids, avoiding iterations. It is worth noting that the extreme score procedure establishes the number of groups and their centroids for K-means clustering which is applied to the RPCs of the 500 hPa geopotential. Hereafter, CT+ and CT– denote the circulation types deriving from positive and negative phases of the RPCs, respectively.
The relationship between CTs and precipitation occurrences was assessed using the Performance Index (PI; Zhang et al., 1997). PI quantifies the relevance of each CT to the occurrence of precipitation, by comparing the daily mean precipitation of a given circulation type, i, with the climatological daily mean precipitation:
where ni is the number of days of pattern i, Ri is the total amount of precipitation falling during those days, and R is the precipitation total received in the entire period of n days (with or without the presence of type i). Thus, a PI(i) value for a particular type i is a measure of the relative contribution of that type to total precipitation; a PI(i) much higher than one indicates that type i has an important contribution to total precipitation.
Results and Discussion
CT1– is characterized by a weak ridge over Turkey that generates westerly winds over Iran, especially in the northern parts. In contrast, CT1+ is characterized by a cyclonic curvature over the Mediterranean and northern Iran and by an anticyclonic curvature over central-southern Iran, inducing rain-generating southwesterly flows. CT2– shows a broad ridge over the north-eastern Caspian Sea, while CT2+ reveals a widely extended trough over the north-eastern sector. Both types generate westerly flow and marginally affect Iranian weather. The spatial configuration of CT3– is very similar to CT2+ over Iran (prevalently zonal), but the centre of the upper cyclonic curvature migrates from north-eastern Russia to north-eastern Europe. CT3+ depicts a ridge over Eastern Europe and a trough over southern Russia. The flow direction over Iran is largely zonal.
The circulation structure of CT4– consists of a trough over the central Mediterranean Sea and a ridge over western Iran, leading to north-westerly flows over the country. CT4+ is characterized by a deep and large trough located over the eastern Mediterranean Sea and the Red Sea, causing south-westerly flows over Iran. The geopotential gradient is strong, making CT4+ very relevant for Iranian weather. A weak as well as widely extended ridge located over Turkey and the Balkans and a weak, shallow and broad trough over eastern Iran characterize CT5–, which leads to north-westerly flows over the country. CT5+, similarly to CT4+, shows a well-established trough over the eastern Mediterranean Sea, favouring precipitation occurrences over Iran. The latter two types differ from each other in the location and direction of their trough axis. The trough location in CT6– is over Eastern Europe, whereas a ridge can be found over Russia, leading to westerly flows over Iran. Finally, CT6+ is characterized by a weak ridge and a trough located over the Balkans and the Red Sea, respectively. This circulation structure might favour precipitation occurrences over Southern Iran, as it transports maritime air masses from the Red Sea and the Persian Gulf.
Spatial patterns of PI related to CT1- implied that this circulation type does not contribute to provide precipitation amounts above their climatological means over western and northern Iran. Conversely, the PI values higher than 1.5 in central-eastern Iran indicate that CT1- tends to contribute to anomalously high precipitation in this part of Iran. The PI pattern for CT1+ shows that it only contributes to precipitation occurrences over west and north-western Iran, leaving most of the country without precipitation occurrences. With respect to the PI pattern for CT2-, it is evident that this CT has no potential for precipitation occurrences throughout the country. CT2+ also tends to be mostly unfavorable to precipitation occurrences over Iran, despite some weak contributions over southern, eastern and northern parts of the country, not having impact over western Iran.
The PI pattern for CT3- clearly highlights the prevailing dryness of this pattern over the whole of Iran. According to the spatial pattern of PI for CT3+, it can also be concluded that it is an essentially dry regime, though it may weakly contribute to precipitation in northern and eastern parts of Iran, where PI values higher than 1.5 are found. CT4- can be considered the driest regime, as its contributions to precipitation are negligible. Regarding CT4+, PI values are greater than 1.5 throughout the country, highlighting its key role on Iranian precipitation. The PI values of CT4+ are very high over southern Iran, emphasizing its pronounced influence in the coastal areas of the Persian Gulf.