Investigation of different interpolation methods for estimation and zoning of precipitation variables in agricultural lands of Aq-Qalla township for rainfed cropping of autumn cereals

Climate condition is one of the most important factors which affect agricultural production. Amongst climate parameters, rainfall variability is one of the most crucial factors which has important role on agricultural productions because of its effect on available water for plants. The amount of water available to plants strongly depends on the rainy seasons onset, length and rainfall temporal distribution, hence can determine a success or failure of a growth season. One of the most important areas for crops production in Golestan province is the Aq-Qalla township. Most of the areas are used for rainfed farming, such as wheat and barely. In general, agriculture in Aq-Qalla township is dependent on rainfall. Determination of the precipitation spatial variability in these agricultural lands is particularly very importance. Therefore, the objectives of this research were the investigation of different interpolation methods for estimation and zoning of precipitation variables in agricultural lands of Aq Qalla township for rainfed cropping of autumn cereals. Also, this research provides information at local level that could be used by farmers to increase of yield in rainfed condition.
In this research, the spatial distribution of precipitation on agricultural lands of Aq-Qalla township evaluated using 18 method-models from deterministic and geostatistic interpolation such as, Inverse Distance Weighted (IDW), Universal Kriging (UK), Ordinary Kriging (OK), Radial Basis Function (RBF) and Local Polynomial Interpolation (LPI). The performance criteria for evaluation were Root Mean Square Error (RMSE) and General Standard Deviation (GSD) with the Cross Validation methods.
The results showed that Local Polynomial Interpolation (1 and 2 degree) was the best method to estimate precipitation parameters except spring and June precipitation. Because, it had the highest accuracy and the lowest error for estimating of these parameters. Semivariograms analysis indicated that annual and June precipitation were best fitted to Exponential model, also, autumn, spring and May precipitation by Spherical model. Results of final maps showed that amount of annual, autumn and June rainfall were lower in north than south regions. Moreover, the results of this research showed that about 65.62 percentage of agricultural lands of Aq-Qalla had suitable level of annual rainfall (300-400 mm). Also, in this area amount autumn rainfall was 101.11 to 166.08 mm.
The results indicated that same as temperature requirements, the annual and autumn rainfall requirements of current crops such as wheat and barley are provided in this township. But, amount of spring and June rainfall are non-suitable in this area. Wide area in Aq-Qalla township were under 10 mm rainfall in June. This amount is not enough for grain filling.