Evaluating the performance of the reanalyzed ERA-Interim database in temporal-spatial distribution and wind speed trend in eastern Iran

Introduction
Wind is the horizontal displacement of air, under one meter per second of speed. It is a dynamic phenomenon with three main characteristics: intensity, direction, and frequency. Therefore, good knowledge of its characteristics in every area of importance is quite remarkable. For instance, while the impact of global warming on temperature and precipitation at global level over the past decades, have been considered in many studies, little attention has been dedicated to wind speed and its influence on climate change. Wind speed alterations can affect the energy of storms, shipping industries, soil moisture, evaporation, and water resources. It may even influence the evolution of dry and semi-arid environments. Furthermore, a lot of research on wind and meteorology has shown that the performance of wind turbines is sensitive to climate change. Possible changes to future wind regimes have been widely considered under changing weather conditions and due to global warming, the intensity and frequency of wind events are expected to alter at the end of this century.
Materials and Methods
The study area in this research on the eastern strip of Iran included four provinces of Khorasan Razavi, South Khorasan, Kerman, and Sistan and Baluchestan. The study used wind speed data in an altitude of 10 meters at 10 synoptic stations. It was provided with daily statistical information between 1985 and 2015, i.e. 30 years of data. When choosing this station, in addition to proper distribution within the region, an attempt was made to select more stations than the one, affected by the 120-day winds of Sistan. The study also employed 10-meter-high wind speed data of the ERA-Interim version with a resolution of 0.125 × 0.125 degrees on a daily basis between 1980 and 2015. For the study area, as many as 3772 pixels with an inter-pixel distance of about 12.5 km were obtained and to evaluate the performance of simulated data against observational data, several indicators were used from the Root-Mean Square Error (RMSE), Mean Bias Error (MBE), Mean Absolute Error (MAE), and the coefficient of determination (R2). It made use of non-parametric Man-Kendall method to order to investigate the trend of wind speed changes.
Results and Discussion
The ECMWF ERA-Interim version delivers a high and good performance for wind speed. Results showed that the output of the mentioned base in all studied stations was on average between 0.722 and 0.984. RMSE, MBE, and MAE characteristics in Zahedan, Khash, and Saravan stations were below 1 m/s. In other words, the wind speed of ECMWF base in these three stations had the highest performance of all 11 stations studied. The monthly statistical assessment of wind speed in the selected stations in eastern Iran during the statistical period (1985-2015) demonstrated that the average wind speed was 3.56 m/s. The correlation between wind speed with negative altitude and positive longitude was significant at the level of 0.05. Also, the relation between latitude and wind speed showed it to be negative during the cold months of the year and positive during the warm ones. The average wind speed fluctuated significantly during the 30-year statistical period, varying between 2.82 and 4.57 m/s. The minimum and maximum wind speeds occurred in December and July, respectively. The average 30-year wind speed at selected stations in eastern Iran turned out to be 2 m/s. The maximum wind speed in eastern Iran displayed many fluctuations, with autumn showing the lowest statistical value of maximum wind speed. In December, this value was 3.98 m/s. The maximum wind speed increased during all studied months. All studied months proved to be statistically significant, except for January, which though increasing was not remarkable at 0.05 and 0.01 levels. Other wind speed studies showed a significant incremental trend at α = 0.01. The average wind speed in the study area was negative in 7 months (i.e., January, April, May, July, August, October, and December) of the year and positive in 5 months (i.e., February, March, June, September, and November) of the other. The maximum wind speed belonged to January (with 4.42 m/s), February (4.86 m/s), and March (with 5.02 m/s.) The next area in the form of a fertile one in winter, Zabol, was also the center of Iran's borders with Afghanistan, near the borderlines of South Khorasan Province. Here, the wind speed trend was positive at the time of 120-day wind onset (June with 0.79), and negative at the time of its termination (October with -0.15).
Conclusion
The average wind speed in the study area (Khorasan Razavi, South Khorasan, Kerman, and Sistan and Baluchestan Provinces) during the long-term statistical period of 30 years (from 1985 to 2015) was 3.56 m/s. The minimum and maximum wind speeds occurred in July and December, respectively. The reason for this increase in wind speed in July was due to the 120-day wind activity in Sistan, which started in June. The average wind speed in the study area was negative in 7 months (January, April, May, July, August, October, and December) of the year and positive in 5 months (February, March, June, September, and November). Investigating wind speed process via non-parametric Man-Kendall (M-K) test showed that the wind speed trend in eastern Iran in the first month of June (June) 120-day winds showed an increasing trend (Z score of the Man-Kendall test 0.795), while in the last month (October) it decreased (-0.1152). Also, in July, when the wind speed was maximum, the average trend in the study area, having a Z score of 0.242, began declining. Pearson correlation test showed that the relation between wind speed and topography in the study area was statistically significant at 0.05. In contrast, the relation between longitude and wind speed was remarkable in all studied moles at an alpha level of 0.05, while neither the longitude nor the altitude in the study area did not show a uniform relation between latitude and wind speed. While this relation was exactly so during the warm months of the year, it was vice versa during the cold ones. In October alone, the relation between wind speed and latitude was not significant.