Investigating the homogeneity of temperature and precipitation data using statistical and statistical-climatic approaches

Document Type : Research Paper

Authors

• Majid Cheraghalizadeh ¹
• Ali Khalili ¹
• Somayeh Hejabi ²
• Javad Bazrafshan ³

¹ Department of Irrigation and Reclamation Engineering, Faculty of Agriculture, College of Agriculture and Natural Resources, University of Tehran, Karaj, Iran

² Department of Water Engineering, Faculty of Agriculture, Urmia University

³ Department of Irrigation and Reclamation Engineering,, Faculty of Agriculture, College of Agriculture and Natural Resources,, University of Tehran,, Karaj,, Iran

Abstract

Keywords

• Adjacent station
• Iran
• statistical-climate
• climate signal
• homogeneity

Main Subjects

• ...

Investigating the homogeneity of temperature and precipitation data using statistical and statistical-climatic approaches

EXTENDED ABSTRACT

Introduction

Managing and policymaking in studies pertaining to natural hazards, such as drought and floods, are significantly influenced by the quality of the data under examination. Consequently, the paramount consideration in addressing management issues lies in ensuring the reliability and quality of the requisite data. The initial step in ensuring the quality control of the data under scrutiny involves verifying data homogeneity. Discontinuities within homogenous time series can be attributed to climatic factors. In contrast, heterogeneous time series exhibit discontinuities due to non-climatic factors, thereby amplifying the level of uncertainty in the findings. The accurate interpretation and reduction of uncertainty in such studies necessitate a meticulous assessment of data quality.

Methodology

This study aimed to examine the homogeneity of temperature variables (average, minimum, and maximum) as well as precipitation within the timeframe spanning from 1990 to 2019. This investigation was carried out using both statistical and statistical-climatic approaches across a dataset comprising 70 synoptic stations in Iran. The statistical approach employed a battery of homogeneity tests, which encompassed Petit homogeneity (PT), homogeneity of cumulative deviations (CDT), Worsley likelihood homogeneity (WLR), standard normal homogeneity (SNHT), and a statistical-climatic approach based on PT in conjunction with data from adjacent stations.

Results and Discussion

The results of the homogeneity tests conducted through the statistical approach revealed that the most frequent breaking year for temperature and precipitation variables was 1997 and 2006, respectively. Significantly, over 90% of the breaking years for temperature data (including minimum, maximum, and average) were deemed statistically significant. Specifically, the highest frequency of breaking years for minimum and maximum temperature occurred in 1997, 2008, and 2000, while for average temperature, it was in 1997, 1994, and 2009. In contrast, less than 15% of the breaking years for precipitation were considered significant, with the highest frequency of breaking years occurring in 2006, 1999, and 1997. Regarding temperature, it was found that 4.3% of stations exhibited homogeneity in minimum temperature, while 92.9% displayed heterogeneity. Other stations were categorized as doubtful. For average temperature, 10% were homogeneous, and 90% were heterogeneous. The analysis of precipitation indicated that 1.87% of stations were homogeneous, 11.4% were heterogeneous, and 1.5% fell into the doubtful category. In the homogeneity investigation based on the statistical-climatic approach utilizing adjacent stations, it was determined that 10% of minimum temperature stations (17.1% of maximum temperature stations) were homogeneous, 64.3% of minimum temperature stations (72.9% of maximum temperature stations) were conditionally homogeneous, and 25.7% of minimum temperature stations (10.0% of maximum temperature stations) were heterogeneous. For average temperature, 11.4% were homogeneous, 91.4% were conditionally homogeneous, and 14.3% were heterogeneous. Similarly, for precipitation, 74.3% were conditionally homogeneous, and 8.6% were heterogeneous.

Conclusion

This study aimed to assess the homogeneity of annual temperature and precipitation data using both statistical and statistical-climatic approaches, with no prior knowledge of metadata or climatic factors. In the homogeneity assessment based on the statistical approach, it was observed that over 90% of the temperature data and precipitation data were likely to be heterogeneous, which could potentially impact data-driven research. Therefore, it is imperative to investigate both climatic and non-climatic factors during the studied period to better understand the data's characteristics. Conversely, recognizing that climatic signals can influence broad geographical areas, it becomes feasible to attribute similar breaking years in a region to climatic factors, irrespective of their specific causes. In such instances, heterogeneity can be regarded as conditional homogeneity, with the heterogeneity factor considered as part of the climatic norm. The results obtained from the statistical-climatic approach indicated that 75% of the heterogeneous temperature data and 100% of the precipitation data could potentially be treated as conditionally homogeneous. However, it is advisable to conduct further in-depth investigations into the statistical-climatic approach to ensure its robustness and reliability.