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

In meteorological and hydrological studies, the initial and crucial step is to assess the quality of the data under investigation. In cases where long-term metadata and prior knowledge regarding data quality are lacking, the reliability of data analysis results becomes questionable across various applications. Therefore, this research aims to conduct a comparative analysis and investigation of the homogeneity of two significant atmospheric variables: temperature (including average, minimum, and maximum) and precipitation. This assessment is carried out using both statistical and statistical-climatic approaches, covering the period from 1990 to 2019 and involving data from 70 synoptic stations in Iran. The findings obtained through absolute statistical approaches revealed that the majority of breaking years for temperature and precipitation variables occurred during the latter half of the 1990s, with a particular emphasis on the year 1997 for temperature and 2006 for precipitation (and 1999 for precipitation, respectively). In general, it was observed that more than 90% of the breaking years for temperature and precipitation variables exhibited heterogeneity, significantly limiting the utility of the data in various applications. However, it is worth noting that understanding the influence of climatic signals on a wide geographical area allows for the attribution of similar breaking years in a region to climatic factors, regardless of their specific causes. This perspective allows for the consideration of heterogeneity as conditional homogeneity, with the heterogeneity factor being seen as part of the climatic norm. The results from this study demonstrate that by employing a statistical-climatic approach based on the concept of adjacent stations (nearest neighbors) to check data homogeneity, it is possible to consider 75% (100%) of the heterogeneous temperature (precipitation) variables as conditionally homogeneous. This approach helps alleviate the limitations associated with using heterogeneous data. Nevertheless, it is advisable to conduct further thorough investigations into the statistical-climatic approach to ensure its robustness and reliability.