Damage evaluation of soybean chilling injury based on Google Earth Engine (GEE) and crop modelling

Abstract

Frequent chilling injury has serious impacts on national food security and in northeastern China heavily affects grain yields. Timely and accurate measures are desirable for assessing associated large-scale impacts and are prerequisites to disaster reduction. Therefore, we propose a novel means to efficiently assess the impacts of chilling injury on soybean. Specific chilling injury events were diagnosed in 1989, 1995, 2003, 2009, and 2018 in Oroqen community. In total, 512 combinations scenarios were established using the localized CROPGRO-Soybean model. Furthermore, we determined the maximum wide dynamic vegetation index (WDRVI) and corresponding date of critical windows of the early and late growing seasons using the GEE (Google Earth Engine) platform, then constructed 1600 cold vulnerability models on CDD (Cold Degree Days), the simulated LAI (Leaf Area Index) and yields from the CROPGRO-Soybean model. Finally, we calculated pixel yields losses according to the corresponding vulnerability models. The findings show that simulated historical yield losses in 1989, 1995, 2003 and 2009 were measured at 9.6%, 29.8%, 50.5%, and 15.7%, respectively, closely (all errors are within one standard deviation) reflecting actual losses (6.4%, 39.2%, 47.7%, and 13.2%, respectively). The above proposed method was applied to evaluate the yield loss for 2018 at the pixel scale. Specifically, a sentinel-2A image was used for 10-m high precision yield mapping, and the estimated losses were found to characterize the actual yield losses from 2018 cold events. The results highlight that the proposed method can efficiently and accurately assess the effects of chilling injury on soybean crops.