Prediction of temperature using SDSM multiple linear models (Case study: Hoor al-Azim and Miangaran wetlands)

Introduction: One of the consequences of the climate change phenomenon in Khuzestan province is the drying up of a large part of the wetlands of this province, including Miangaran and Hoor al-Azim, which has caused dust storms due to their continued drying in recent years. In this regard, this research aims to predict climate changes in the area of Miangaran and Hoor al-Azim wetlands by SDSM statistical microscale model based on HadCM3 B2 and A2 climate scenarios. Considering the specific conditions of the region and the fact that few studies have been done regarding temperature change in these areas, knowing the state of temperature change can help better management of resources, especially water resources management.
Materials and methods: These parameters include the average sea level pressure, the geopotential height of the surface of 850 hectopascals and the average temperature at a height of two meters. For this purpose, by using the daily data of average temperature, minimum and maximum temperature in the synoptic stations of Izeh and Bostan as the closest stations to Hoor al-Azim and Miangaran wetlands in the periods of 2010-2039, 2040-2069 and 2070-2099, predictions were made and Then a comparison was made with the base period (1961-1990). To check the performance of the models and compare them, two drawing methods were used conventional statistical criteria of PBISE, NSE, MAE, RMSE and R2.
Results and discussion: The results of the simulation of the HadCM3 model along with observational data for the Izeh station, modeled annual average temperature data was 18.47°C and for the Bostan station, modeled annual data average temperature was 19.10°C. As can be seen, both stations had a higher average temperature in the base period, and the maximum temperature in the Miangaran wetland is much higher than Hoor al-Azim wetland in the base period. The minimum temperature results show that the difference between two stations is small and minor. The results showed that in both stations in scenario A2, the average temperature had significant cycles with return periods of 1.2 years and the lowest significant cycles for the two stations were in return periods of 2.3 and 1.3 years, respectively. In scenario B2, the average temperature in two stations has significant cycles with return periods of 7.5 years and the lowest significant cycle with return period of 1.2 years. The results of examining the cycles in the studied areas indicate that in the A2 scenario, certain climatic conditions in the area have short-term return periods. In two stations in scenario B2, in the upper percentile of the significance cycle, the mean temperature is at the confidence level (97.5%) in the 30-year return period, and in the lower percentile of the significance cycle, the mean temperature is at the confidence level (2.5%) in the return period. 1 year was obtained in both scenarios, which indicates a wider range of the return period and a higher probability of extreme temperature events in the B2 scenario.
In the comparison of the two stations, it can be seen that in scenario A2, the average temperature of the Miangaran wetland is 1.02 °C and Hoor al-Azim wetland is 1.08 °C. In scenario A2, Miangaran wetland has a higher average and maximum temperature, while Hoor al-Azim wetland has a higher minimum temperature increase. While in scenario B2, the average temperature of Miangaran lagoon is 1.47 °C and Hoor al-Azim wetland is 5.12 °C.
Conclusion: The results of the data analysis in the future observation and simulation periods with scenarios B2 and A2 show an increase in the mean, minimum, and maximum average temperature in the future simulation periods compared to the base period in Izeh and Bostan stations so in Bostan station, the average minimum and maximum annual temperature also increased in the third period compared to the base period. And in both scenarios, due to the increase in temperature, the drying process of both wetlands will continue.