Document Type : Research Paper

Authors

• Zahra Heydari monfared ¹
• Seyed Hossein Mirmousavi ²
• Hossein Asakereh ³
• Koohzad Raisipour ⁴

¹ Doctoral student of hydrology and meteorology, Zanjan University

² Associate professor , Faculty member of Zanjan University

³ Professor of Zanjan University

⁴ Assistant professor of Zanjan University

Abstract

Extended Abstract
Introduction: Snow-cover changes and related phenomena (especially depth, snow water equivalent and snow density) have a fundamental role in mountainous environments and strongly affect water availability in downstream areas. In this way, the importance of correct and appropriate analysis is more visible. Due to the fact that most of the rainfall falls in the form of snow in mountainous areas, the management of snow resources in these areas is very important, and knowing the different aspects of variability and geographical patterns governing the phenomenon of snow is a scientific and practical need. It is considered special in water resources and in the agricultural sector. Thus, in the current research, the spatio-temporal patterns governing the annual average of snow density in different decades and the difference of each of the decades compared to the entire time period have been estimated and analyzed using spatial statistics methods.
Materials & Methods: The studied area with an area of ​​about 151,771.91 square kilometers is located between 34°44' to 39°25' north latitude from the equator and 44°3' to 49°52' east longitude from the Greenwich meridian. In order to investigate the spatial autocorrelation changes of the average snow density in northwest Iran during the years 1982-2022 from the data obtained from the database of the European Center for Medium-Range Atmospheric Forecasting ECMWF⁴/ ERA5 based on daily data, and to identify and understand the spatial patterns of density Barf, based on statistical and graphic models have been used in the geographic information system environment. In the study of temporal-spatial changes of the average snow density of the region in different time periods including 4 decades ((1982-1992), (1992-2002), (2002-2012), (2012-2022)) and the whole period of 41 years (2022) -1982)), general Moran's I and Getis-Ord Gi* statistics were used. Also, in the current research, in order to investigate the effect of changes in  Extreme  snow precipitation on the amount of snow density in the northwest region, it has been done to determine the snow threshold. In order to estimate snow drift, a threshold was defined. Since the station snowfall amount data has a high dispersion, values ​​above the mean cannot be accurate for defining the threshold of freezing snow. In this way, the 99th percentile index has been used to determine the snow threshold.
Results & Discussion: The aim of the current research is to investigate the spatial autocorrelation changes of the annual mean snow density in the northwest of Iran. For this purpose, the annual snow density data during the statistical period of 1982-2022 was obtained from the ECMWF/EAR5 database with a resolution of 0.25 x 0.25 degrees, and then divided into four ten-year periods. In order to analyze spatial autocorrelation changes, global Moran indices and hot spot analysis (Gettys-RDJ) were used at the significance level of 90, 95 and 99%. Also, in order to investigate the effect of extreme precipitation on changes in the level of snow density, the 99th percentile statistical index was used, and based on this index, the freezing threshold of each synoptic station in the region was determined during the last decade (2012-2022) and the interval the entire statistical period (1982-2002) was carried out. The results of the present research showed that in the studied area, snow density has spatial autocorrelation and a strong cluster pattern. With a density threshold less than 0.10 kg/m³, from the first decade to the end of the fourth decade, the area (number of pixels) and the amount of snow density in the northwest have decreased. The results of the analysis of the changes in precipitation in the 99th percentile showed that the amount of this type of precipitation has increased significantly during the last decade of the study, and this has caused the snow density to increase relatively in the last decade compared to the first to third decades. However, in general, the amount of snow density in the entire northwest area has significantly decreased during the last four decades.
Conclusion: The evaluation of the temporal changes of snow density also strengthened the hypothesis of the occurrence of freezing snow precipitation leading to an increase in snow density in the months of cold seasons during the last decade. This point was confirmed by examining the statistical index of the 99th percentile of snowy days of each synoptic station in the region during the last decade (2009-2018) compared to the entire period of station statistics (2000-2018). The results of the analysis of the changes in precipitation in the 99th percentile showed that the amount of this type of precipitation has increased significantly in the last decade of the study and this has caused the snow density in the last decade to increase relatively compared to the first to third decades. However, in general, the amount of snow density in the entire northwest area has decreased significantly during the last four decades. Moran's statistic was used to explain the pattern governing snow density in northwest Iran. The results of Moran's index about the annual average of snow density showed that the values ​​related to different time periods have a positive coefficient and are close to one, which indicates that the snow density data has spatial autocorrelation and has a cluster pattern. Also, the results of standard Z score and P-value confirmed the cluster significance of the spatial distribution of snow density in the northwest. Finally, the analysis of hot spots has been a clear confirmation of the continuation of concentration and clustering of snow density in northwest Iran in space with the increase of the time period, which mountainous areas have the first rank in the formation of hot clusters with a probability of 99%. have given.

Keywords

• Snow density
• Spatial statistics
• Spatial autocorrelation
• Moran index
• hot spots
• Cluster pattern
• Northwest

Main Subjects

• Geographic Data