Spatio-temporal variations and drought of spring maize in Northeast China between 2002 and 2020.

Abstract

A lot of spring maize is grown in Northeast China (Liaoning, Jilin, and Heilongjiang), an area that is highly susceptible to drought. Here, remote sensing indexes from 2002 to 2020 were studied using the 8-day surface reflectance and land surface temperature of Moderate-resolution Imaging Spectroradiometer data. Spring maize distribution was extracted using a decision tree classification, and the results were compared to the known distribution based on field investigation data and published statistics. The results showed that mixed pixels of spring maize and soybeans had limited influence on the study of spatio-temporal variations of spring maize, and the error was acceptable. The overall accuracy of verifying the spring maize distribution from 2018 to 2020 was above 85%. The stable, fluctuating, and low-frequency planting areas of spring maize accounted for 11.86%, 17.41%, and 34.86% of the study area, respectively. In 2015, the government directed a reduction of the planting area of spring maize in the "Liandaowan" region of Northeast China. The planting area of spring maize was characterized by a continuous increase before this change (2002-2014), exhibited changes and reductions in response to the change (2015-2017), and exhibited optimization and recovery after this change (2018-2020). Compared with the fluctuating and low-frequency planting areas, moderate and severe droughts were higher in stable planting areas. From 2002 to 2020, the most severe droughts occurred in the expanded planting areas. This rapid and large-scale monitoring of spatio-temporal variations and drought of spring maize provides a foundation for improving grain yield. This method could be easily applied to the study of other regions and combined with high-resolution and hyperspectral satellite data to improve monitoring accuracy.