Analysis of 70 Years Daily Precipitation in Tabriz Using Visual Data Mining Approach

Authors

1 Assoc. Prof., Dept. of Water Engineering, Faculty of Agriculture, University of Tabriz

2 M.Sc., Dept. of Statistics, Faculty of Mathematic, University of Tabriz

3 M.Sc. Student, Dept. of Water Engineering, Faculty of Agriculture, University of Tabriz

Abstract

Abstract
Background and Objectives
Daily precipitation, which is completely stochastic, is one of the basic components of the water cycle and has an important role in the management of surface and ground water resources in terms of quantity and quality. Indispensable element of drought analysis and flood control research is precipitation. The efficient management of surface water resources directly depends on precipitation. The lack of long-term and reliable data has made it difficult to determine precipitation behavior. In this study, we tried to determine the behavior and pattern of precipitation through human-data interaction called visual data mining approach. One of the new approaches that focuses on the use of visualization and graphics in the analysis of complexities in data is visual data mining. Visual data mining can be thought of as a combination of two disciplines, visualization and data mining. Visual data mining is also closely related to computer graphics, multimedia systems, human-computer interaction, pattern recognition, and high-performance computing. The aim of this study is to analyze Tabriz daily precipitation data and discover the patterns in this data with the help of visual data mining approaches. Discovering these patterns and identifying rainfall behavior will help to manage floods on the one hand and droughts on the other.
Methodology
In this study, daily precipitation and temperature data of Tabriz Synoptic Station for the last seventy one years (1951-2021) were used to determine precipitation patterns. Tabriz has a generally cold climate as the center of the eastern Azerbaijan province and is surrounded by mountains. Recently, with the increase in data and software, data mining techniques have also started to attract attention. Data alone cannot mean anything. However, graphs consisting of data can give very meaningful information and messages. Visual data mining approach is an effort to bring data to life with different graphics. In this research, various softwares such as R, ArcGIS and Tableau were used for visualization. In addition, ExcelStat was used to check the accuracy of the data and for statistical tests. R statistical language was used to create the data mining structure and to display the data graphically. Then, different diagrams were drawn using the Tableau program. Finally, the drawn diagrams were evaluated and the final graphics were selected. ArcGIS software was used for spatial analysis and map drawing. Also, multiple linear regression method was used to predict precipitation amount and probability of occurrence.
 
Findings
According to the temperature histogram, the long-term average annual temperature in Tabriz varies between 12 and 13 degrees Celsius. Also, according to the precipitation histogram, precipitation over 5 mm in Tabriz varies between 10 and 20 days per year on average. The results obtained showed that there has been a change in precipitation patterns in the last 5 years (2017-2021). Although annual precipitation during these five years is above the 71-year average, it is still below the golden precipitation period (1961 to 1970). The results showed that the intensity of spring rains decreased significantly in Tabriz during the period 2006-2021. However, from 1971 to 1980 and from 2001 to 2010, it was observed that the spring rains were more in terms of both precipitation values and precipitation intensity. However, in the following periods from 2011 to 2021, both precipitation intensity and precipitation values decreased. The results showed that most of the spring precipitation in 1981 was 276.3 mm, making up 73.26% of the total precipitation for that year. According to the findings, precipitation has started to decrease in the spring season in recent periods. Decreased dry grain yield in the Azerbaijan region may be affected by decreased spring precipitation. As a result of this decrease, it is expected that the agricultural economy in the study area will be negatively affected.
Conclusion
In this study, as a first, daily precipitation in Tabriz was investigated with visualized data mining techniques. Thus, interesting findings were obtained with the help of different graphics. The results showed an increase in precipitation in the last 5-6 years. It has also been proven that the temperature behavior during this period shows an increase in average temperature, a confirmation of the increase in global temperature. Although the results of this research showed that visualized data mining is successful in precipitation analysis, it is recommended to conduct more comprehensive studies in this field in the future.

Keywords


Arthi Rani B, Manikandan N and Maragatham N, 2014. Trend analysis of rainfall and frequency of rainy days over Coimbatore. MAUSAM Journal 65 (3): 379-384.
Asakereh H and Akbarzadeh Y, 2017. Simulation of temperature and precipitation changes of Tabriz Synoptic Station using statistical downscaling and canesm2 climate change model output. Journal of Geography and Environmental Hazards 21: 153-174. (In Persian with English abstract)
Daniel A, Christian P and Mike S, 2003. Visual data mining of large spatial data sets. Lecture Notes Computer Science 2822: 201-215.
Darand M and Ebrahimi B, 2019. Temporal-spatial analysis of changes in waiting time for rainfall in Kurdistan province. Journal of Water Resources Engineering 11: 17-30. (In Persian with English abstract)
Ferdosi F and Dinpashoh Y, 2019. Investigation of temporal distribution of daily rainfall using the normalized rainfall curves (NRC) (case study: Tabriz Station). Journal of Water and Soil Science 2)29(: 1-14. (In Persian with English abstract)
Garcia JRM, Monteiro AMV and Santos RDC, 2012. Visual data mining for identification of patterns and outliers in weather stations’ data. Pp. 245-252. International Conference on Intelligent Data Engineering and Automated Learning. 29-31 August, Natal, Brazil.
Gentilucci M, Barbieri M, Lee L and Zardi D, 2019. Analysis of rainfall trends and extreme precipitation in the middle adriatic side, marche region (Central Italy). Water 11 (9). doi:10.3390/w11091948.
Ghorbani MA, Mahmoud Alilou S, Javidan S and Raghavendra S, 2021. Assessment of spatio-temporal variability of rainfall and mean air temperature over Ardabil province, Iran. SN Applied Sciences 3 (728). https://doi.org/10.1007/s42452-021-04698-y.
Han J and Pei J, 2012. Data Mining Trends and Research Frontiers. Data Mining (3rd Edition) Pp. 585-631. https://doi.org/10.1016/B978-0-12-381479-1.00013-7.
Iranpour F and Zohreh Wandi H, 2019. Analysis of daily rainfall in Tabriz to study the probability of frequencies and the persistence of dry and wet days. Journal of Climate Research 36: 91-105. (In Persian with English abstract)
Jahanbakhsh S, Abtahi V, Ghorbani MA, Tadayoni M and Valayi A, 2015. Temporal and spatial distribution of rainfall in Tabriz county using hierarchical cluster analysis. Quarterly Journal of Geographical Space 50 (15): 59-81. (In Persian with English abstract)
Jana C, Sharma C, Alam N, Mishra PK, Dubey S and Kumar R, 2016. Trend analysis of rainfall and rainy days of Agra in Northern India. International Journal of Agricultural and Statistical Sciences 12 (1): 263-270.
Keim DA, 2002. Information visualization and visual data mining. IEEE Transactions on Visualization and Computer Graphics 8 (1): 1-8. doi: 10.1109/2945.981847.
Keim D and Ankerst M, 2005. Visual data mining techniques. Visualization Handbook. Pp. 831-843. https://doi.org/10.1016/B978-012387582-2/50045-9.
Kopanakis I, Pelekis N, Karanikas H and Mavroudkis T, 2005. Visual techniques for the interpretation of data mining outcomes. Pp. 25-35. Panhellenic Conference on Informatics. 11-13 November, Volas, Greece.
Ostertagova E, 2012. Modelling using polynomial regression. Procedia Engineering 48: 500-506.
Pawar P, Rawat U, Yadav A, Rajput A, Vasht D and Nema S, 2020. Long term trend analysis of rainfall, rainy days and drought for Sindh River Basin, Madhya Pradesh, India. International Journal of Current Microbiology and Applied Sciences 9 (12): 2738-2749.
Simoff SJ, Bohlen MH and Mazeika A, 2008. Visual Data Mining: An Introduction and Overview. Visual Data Mining. Lecture Notes in Computer Science, vol 4404. Springer, Berlin, Heidelberg. 1–12.
Vagh Y, 2012. The application of a visual data mining framework to determine soil, climate and land use relationships. Procedia Engineering 32: 299-306.
Ward M, Peng W and Wang X, 2004.  Hierarchical visual data mining for large-scale data. Computational Statistics 19: 147-158.
Yacoub E and Tayfur G, 2020. Spatial and temporal of variation of meteorological drought and precipitation trend analysis over whole Mauritania. Journal of African Earth Sciences 163 (1): 1-12. https://doi.org/10.1016/j.jafrearsci.2020.103761.