Between Fixed and Extending Means for Quantifying Historical Drought Events

Abstract

Drought analysis has become an essential task, due to the impacts of drought on global food security and all aspects of human life. However, drought monitoring does not follow a specific formula but is more a process, thus it is important to examine the details accounted by using different approaches. The Standardized Precipitation Index (SPI), the most widely used for quantifying drought, represents drought intensity for a given time window based on a long-term precipitation mean for a time series. This paper briefly discusses the use of “fixed” or “extending” means; a meteorologist or a webapp recording drought in real time and documenting values in a fixed archive may be using an extending mean, as their long-term mean may change each subsequent year. Meanwhile, someone calculating SPI for historical years may consider that the mean precipitation time series up to his current time as the long term “fixed” mean. In this paper, the 3-month SPI for Al-Mafraq in Northern Jordan was calculated for the period 2011 – 2019, once using a fixed mean, and once using an extending mean method. The results show that no considerable difference in values can be detected (mean absolute error = 2%), but different drought categories may result from the two methods, particularly in the beginning of the time series. This paper strongly suggests that all details involved in calculating drought indices should be communicated when sharing drought-related data, and some terminology should be agreed upon. This paper also presents some suggestions when considering an approach or software for drought analysis and quantifying historical drought.