Investigation of Spatial Distribution and Changes in Temporal Characteristics of Winter Season in the Lake Urmia Basin

Introduction

In climate studies based on astronomical methods, static temporal and spatial definitions of seasons are used. However, seasons change with time and space. Changes in the length and timing of seasons have major impacts on nature and human society. For example, shorter winters and reduced snow cover combined with an earlier onset of spring lead to earlier and weaker spring floods and increased river discharge in winter. Changes in climatic parameters, especially snowfall and snow storage, significantly affect the seasonal water balance and the precipitation-runoff process. Furthermore, these changes are predicted to continue in the future due to further climate change. The large range of elevation changes in the Lake Urmia basin has caused the distribution of snow cover in the basin to be heterogeneous. Global warming caused by greenhouse gas emissions can affect the temporal and spatial distribution of snow cover in the Lake Urmia basin, both through changes in the timing and length of the winter season and through increased evapotranspiration, which has the largest contribution to controlling snow cover area. In this regard, this study aims to evaluate and compare the characteristics of the start, end, and length of the winter season in two recent 30-year periods (1964-1993 and 1994-2023) using a winter season definition method based on daily air temperature data. It is worth noting that the method used is based on the inclusion of spatial heterogeneity of winter season characteristics and allows for the investigation of the relationship between winter season characteristics and altitude.



Materials and Methods

In this study, data from the ERA5-Land reanalysis database were used. Daily ERA5-land data were extracted by coding in Google Earth Engine for the Lake Urmia basin and for the two recent 30-year periods, including 1964-1993 and 1994-2023. To determine the winter season, instead of calendar winter, a cell-based definition was used. In this method, winter was defined based on the 25th percentile of daily temperature (T25) based on each of the two periods. With this method, a set of days with T25 less than zero degrees Celsius was defined as winter. Therefore, the beginning and end of the winter season were defined as the first and last day of the year with T25 less than zero degrees Celsius. The start and end of the winter season were reported in terms of the Julian day-number (starting from January 1), and the length of the winter season (in days) was calculated as the interval from the start to the end of the winter season. Next, the spatial distribution and changes in the day-number of the start, the day-number of the end, and the length of the winter season were compared for the two periods. Also, the correlation of the start, end, and length of the winter season with altitude was examined. For this purpose, global SRTM digital elevation model data with a spatial resolution of 30 meters was used. The elevation corresponding to the coordinates of each ERA5-land grid cell within the basin was extracted using ArcGIS software.



Results and Discussion

The percentage of cells with an increase in the day-number of the start of the winter (46.11%) is very close to the percentage of cells with a decrease (47.96%), and the average changes in cells with increasing or decreasing changes are, respectively, 7 and -7.5 days. The spatial distribution of the difference in the day-number of start of the winter is uniform at the basin and has a median of zero (meaning no change in the start of the winter).

Regarding the end of the winter, the differences in most cells (82.45%) are negative, and the average changes in the day-number of the end of the winter in cells with increasing or decreasing changes are, respectively, 7.5 and -15 days. In general, the day-number of the end of the winter has a median of -11 days, which indicates an earlier end of the winter in the second period (1994-2023).

The length of the second period winter in most cells (73.92%) is shorter than the first period. The average change in winter length in cells with increasing or decreasing changes is -17.1 and -17.4 days, respectively. The median change in winter length is also similar to the end of winter, -11 days, indicating a decrease in winter length in the second period.

In both periods, the start, the end and the length of winter show significant negative, positive and positive correlations with altitude, respectively, and the slope of change is slightly steeper in the second period.



Conclusion

The spatial distribution of the start, end, and length of the winter season is non-uniform and follows the topography of the Lake Urmia basin; so that with increasing altitude, the start of the winter season occurs earlier and the end of the winter season occurs later. Therefore, the length of the winter season is longer at higher altitudes. The median change in the length of the winter season is -11 days, which indicates a decrease in the length of the winter season in the second period (1994-2023) compared to the first period (1964-1993). The sensitivity of the characteristics of the winter season to global warming is non-uniform at the basin and the rate of change is greater at lower altitudes. As the trend of increasing air temperature continues, it is predicted that the decrease in the length of the winter season will be even more severe in the coming decades, because according to previous studies under climate change scenarios, the length of the winter season will decrease significantly and reach less than 2 months in the middle latitudes of the Northern Hemisphere. Since global warming can affect the temporal and spatial distribution of snow cover in the Lake Urmia basin, it is recommended to predict changes in winter characteristics in the Lake Urmia basin under climate change scenarios for future studies.