Statistical Forecast of Daily Maximum Air Temperature in Arid Areas at Summertime

Abstract

Based on historical observations of summers for the period from 2004 to 2018 with a focus on daily maximum and minimum air temperatures and wind speed recorded at 0600 GMT, a non-linear regression hypothesis is developed for forecasting daily maximum air temperature (Tmax) in arid areas such as Baghdad International airport station, which has a hot climate with no cloud cover or rain. Observations with dust storm events were excluded, thus this hypothesis could be used to predict daily Tmax on any day during summers characterized by fair weather. Using mean annual daily temperature range, daily minimum temperature, and the trend of maximum temperature with wind speed, Tmax was forecasted and then compared to those recorded by meteorological instruments. To improve the accuracy of the hypothesis, daily forecast errors, bias, and mean absolute error were analyzed to detect their characteristics through calculating relative frequencies of occurrence. At the end of this analysis, a value of (-0.45ºC) was added to the hypothesis as a bias term.

Keywords

• bias
• daily temperature range
• maximum air temperature
• mean absolute error
• non-linear regression equation