Prediction of Meteorological Drought with Different Methods in Dongsheng of Inner Mongolia, China

Abstract

Drought is one common natural disaster throughout the world. Meteorological drought is a moisture deficiency status caused by the unbalance between precipitation and evaporation. The intensity of drought is usually reflected by the degree of rainfall deficiency. Dongsheng District of Ordos, Inner Mongolia, China belongs to the continental climate of the temperate zone and is prone to climate changes. Thus, studying the effects of regional climate changes on ecological restoration and agricultural production is of high significance.  In this study, meteorological data from 1988 to 2018 in this region were collected, and then the major influence factors and occurring rules of meteorological drought in these 31 years were analyzed. Two drought indices suitable for this region were selected. Afterwards, drought prediction models were constructed and used to forecast the occurrence of drought in the future 5 years. (1) Correlation analysis of meteorological factors showed 4 of 8 factors were highly correlated, including relative humidity, precipitation temperature and evaporation, and the correlation coefficients were 0.751, 0.747 and 0.651. Air temperature and ground temperature were positively correlated and gradually rose year by year. The sunshine duration varied largely year by year and was correlated negatively with precipitation and positively with evaporation. The precipitation was negatively correlated with evaporation, but the precipitation did not largely vary from year to year. (2) Prediction models combining back-propagation neural networks and PA were of high precision in prediction. It was predicted the PAs of 2020 and 2022 were as low as -19.069 and -18.951 respectively, and thus gentle drought will occur in these two years.