تحلیل روندِ زمانی و مکانی الگوهای گردشی جو و اثرات آن بر بارش‌های ایران در دو دهۀ اخیر

A B S T R A C T
In this research, the daily data of geopotential height of 500 hectopascals (hPa) with a spatial resolution of 1 degree from the ECMWF database for Southwest Asia and rainfall station data from the National Meteorological Organization (1979 to 2018) have been exerted. The technique used the principal component analysis and cluster analysis. With these analyses, nine circulation patterns were identified. The changes in the patterns were tested at the 95% significance level by the non-parametric Mann-Kendall test, and Sen's slope estimator was exerted to estimate the number of changes. The significance test of the trend for the winter patterns in Iran's rainy season revealed the significant trend of increasing the height of the geopotential, which has led to a decrease in the pressure gradient and a decrease in instability, and finally, a weakening of the winter precipitation patterns. Significant positive trends of geopotential height showed the continuation of these conditions for summer patterns (increasing stability, decreasing rotation, and decreasing precipitation). Of the nine known patterns, only one seasonal pattern showed a significant negative trend in the country. This pattern, with a slight increase in rainfall, indicates the formation of unstable conditions, which can lead to moderate-season rains if moisture is available. The findings showed that a rainy winter pattern had been eliminated in the last two decades, and a summer pattern had appeared instead.
Extended Abstract
Introduction
Atmospheric circulation patterns play an essential role in the emergence of environmental phenomena, which is why the classification of weather systems is one of the main goals of synoptic climatology. With the advent of computers and advanced mathematical algorithms, such as principal component analysis (PCA) and cluster analysis (CA) methods, as well as the availability of digital data, quantitative methods replaced manual methods. Most methods used and discussed for classifying circulation patterns are based on using multivariate statistics, principal component analysis, and clustering techniques. This research uses the same method to classify atmospheric circulation patterns. Due to a large amount of data, MATLAB software was used in this research.
 
Methodology
The statistical population of this research includes the rainfall station data of the National Meteorological Organization from 1979 to 2018, which have been converted into grid data (2491 cells) with a resolution of 0.25 degrees using the kriging interpolation technique. For typification of daily data of geopotential height level of 500 hectopascals (hPa) for the frame (coordinates) zero to seventy degrees east longitude and ten to sixty degrees north latitude from ECMWF European Center for Medium-term Atmospheric Forecasting, ERA-INTERIM project from 1/1/1979 to 12/31/2018 has been used for 14610 days. The data were divided into two 20-year periods for a two-decade comparison. This framework was considered significant enough to represent the circulation patterns affecting Iran's climate properly.
Finally, the data matrix was prepared with two matrices with dimensions of 3621 x 7305. Then principal component analysis was performed on these two matrices. The purpose of this analysis is, on the one hand, to reduce the amount of data and, on the other hand, to classify and identify the most important patterns and changes in geopotential height of 500 hectopascals (hPa) in the last two decades. Twelve components of the S matrix with a level of 500 hectopascals (hPa) were used as the required input for the following classification step to identify the types of air and classify them. Then, nine patterns or weather types were identified by cluster analysis. With the help of the Mann-Kendall test and Sen’s slope estimator, pattern changes were done on time and places (pixels).
 
Results and discussion
The correlation coefficient parameter was used to identify similar patterns in two periods. In this way, three winter patterns, three temperate season patterns, and two summer patterns were determined. Pattern 3 from the first period is a winter pattern, and pattern seven from the second period is a pattern with the features of the warm season, and no suitable pair was identified. These two patterns had the lowest correlation coefficient with each other. It is seen that the CTA3 pattern, a winter pattern with heavy rainfall, was removed in the second period, and the CTB7 pattern, a spring-summer pattern with little precipitation, was born instead.
The Mann-Kendall trend test on the patterns did not show a negative trend in the time series for any pattern. Two pairs of winter patterns have a significant positive trend, and pattern number 3 was removed. Two pairs of the temperate season pattern and two pairs of the summer pattern showed a significant positive trend, and the seven summer patterns appeared in the second period.
The trend test on the pixels of the region for the pattern of one winter showed all of Southwest Asia with significant positive trends, which indicates the weakening of this pattern with warmer winters. The second winter pattern in the country's eastern half shows the weakening of the second cold season with wide positive trends. Another noteworthy point is the significant negative trends for the pair of moderate CTA5B4 patterns significantly and widely over our country, which can lead to rain if other conditions are available.
The two pairs of the summer pattern have covered almost the same range in terms of the significance of the trend and its values. Significant positive trends (increase in geopotential height) for summer patterns provide conditions for increasing stability, reducing rotation, and reducing precipitation.
The conducted analyses show that under the influence of climate change, the rule of a hotter and drier climate in our country in the last two decades is quite evident. The expansion of low rainfall areas can be clearly seen for all patterns. The comparison of the rainfall maps of the country related to the pair of winter patterns PA1, PB1, and PA2, PB2, and PA9, PB3 shows that in addition to the decrease in the rainfall of these patterns, their spatial distribution has also undergone significant changes. The core of the maximum rainfall from the country's west to the southwest side has been moved.
 
Conclusion
A side-by-side comparison of the models showed significant changes in the models. The patterns associated with high-altitude and ridge settlements on all or a large part of Iran are more frequent, consistent with Masoudian's research (2006). The significant positive trend in the Sudan and Mediterranean circulation systems, which play an essential role in the rains of our country's winter and autumn seasons, revealed the weakening of these systems in the last two decades. These results are in harmony with the research of Alizadeh (2013) and Darand (2014). Another result of this research is that the patterns of Iran's rainy seasons (winter and autumn) have weakened significantly in the past two decades. Significant positive high-altitude trends for summer patterns showed increasing stability and strengthening of these patterns. Significant positive high-altitude trends for summer patterns showed increasing stability and strengthening of these patterns. Also, the CTA4B5 transition pattern pair showed significant negative trends over a wide part of the country; the nature of this pattern determined that with the establishment of the CTB5 pattern (the second-period pair) if moisture is available, it can provide the possibility of widespread rains in the country. Correlation coefficients identified two inconsistent patterns. The CTA3 pattern is a winter pattern with heavy rainfall that has not occurred in the last two decades and can be said to have disappeared, and instead, the CTB7 pattern is a summer pattern that has appeared with a frequency of 10.7% in the last two decades.
 
Funding
There is no funding support.
 
Authors’ Contribution
All of the authors approved thecontent of the manuscript and agreed on all aspects of the work.
 
Conflict of Interest
Authors declared no conflict of interest.
 
Acknowledgments
We are grateful to all the scientific consultants of this paper.