Abstract:Currently, it is widely recognized that crop growth stages in Northern arid region of China, especially the Yellow River Basin (YRB), have experienced significant changes due to many factors including climate change. Rasterizing crop growth stages, as an essential input for analysis of crop pattern, is helpful for field crop management as well as early crop production estimation. The Intergovernmental Panel on Climate Change (IPCC 2001) reported that global average surface temperature has increased over the 20th century by 0.6 ℃. Therefore it is necessary to develop a good methodology of rasterizing spring maize growth stages to carry on the further research in changes of crop growth stages under future climate change scenarios with consideration of the quantitative relation between crop growth stages, weather data, and geographic features. Given the above background, the aim of this study is to develop the spatial patterns of spring maize growth stages in the YRB. Multiple-step regression (MSR) is conducted to analyze the relationship between each growth stage of spring maize and their influencing factors, such as longitude, latitude, topography, annual precipitation, annual mean temperature, ≥10℃ accumulated temperature and sunshine hours for year 2000 to 2008. The grid maps of four different growth stages were generated and the accuracy was analyzed through paired sample test with SPSS. The results showed that the maps of sowing date, anthesis stage and harvest stage had the highest accuracy, while elongation stage had satisfactory accuracy. The sowing dates exhibit a delayed trend from southwest to northeast. The other three stages, however, were found to be later from the north and south to central region, and then to the east and west. The grid maps of four crop growth stages described that the time span of sowing date was the shortest with 20 days. The time span of harvest and elongation stages was followed. The anthesis stage was the longest with 50 days because of significant differences in geographic and weather conditions. The best correlation was found between maize sowing dates, longitude, annual precipitation, and ≥10℃ accumulated temperature. Good correlation was also found between elongation stages, longitude, latitude, annual precipitation and≥10℃ accumulated temperature. The correlation between anthesis stage, longitude, annual precipitation, annual mean temperature, ≥10℃ accumulated temperature, and also between harvest stage and longitude, DEM, annual precipitation, annual mean temperature and ≥10℃ accumulated temperature are all good but at less significant level. The MSR approach is proved a robust method to rasterize spring maize growth stages, which enables to develop active adaptation measures for agricultural production under the influence of climate change. The analysis was constrained by the lack of data including spring maize growth data of phonological station in YRB, the data about soil temperature steadily pass the 10℃ in spring maize sowing stage, and spring maize growth data influenced by different varieties, which will contribute to the mitigation of meteorological disaster and suitable maize varieties breeding in YRB.