عنوان مقاله [English]
Air temperature is one the most important variables required for environmental and agricultural studies which are not generally available with sufficient spatial and temporal resolution. Thus, the spatial and temporal disaggregation of properties of the catchment is essential for optimal management of the catchment. The common interpolation functions, including fractal, and regression can produce reasonable results. In this research the interpolation functions based on fractal and periodic regression models were used for modeling and disaggregating temperature datasets for the period of 2007- 2009 at Mashhad and 1980-1982 at Kerman Synoptic stations, respectively. At first, two produced daily temperature from daily datasets. Then data with 5-day and 10-day intervals were used to produce daily temperature. Second, we considered data to be missing at random, and then periodic regression and fractal interpolation were adopted to model daily temperature and then to generate 3-hours temperature. On general results showed similar trends in both climates, and 5-day intervals performed more acceptable, such that determination coefficient for Mashhad and Kerman was 0.98 - 0.77 and 0.98 - 0.82, respectively, while RMSE was between 1.52 - 5.81 and 1.19 - 5.48 anddeg;C, respectively. The intercepts and slopes of regression lines between measured and predicted temperatures were not statistically (5% level of significant) different from 0 and 1, respectively. On the overall, fractal interpolation was better than periodic regression.