Risk analysis using climate-induced yield losses (CIYL) extracted from long-term yield data have been recognized in China, but the research focusing on the time-series characteristics of risk and the circulation signals behind yield losses still remains incomplete. To address these challenges, a case study on winter wheat production in Henan province, north China was conducted by using annual series of yield in 17 cities during 1988–2017 and monthly series of 15 types of large-scale oceanic-atmospheric circulation indices (LOACI). A comprehensive risk assessment method was established by combining the intensity, frequency, and variability of CIYL and principal component analysis (PCA). The results showed that the westernmost Henan was identified as the area of higher-risk. PCA and Mann–Kendall trend tests indicated that the southern, northern, eastern, and western areas in Henan province were classified as having different annual CIYL variations in these four sub-regions; the decreasing trend of CIYL in northern area was the most notable. Since the 2000s, a significant decline in CIYL was found in each sub-region. It should be noted that the key LOACI, which includes Tropical Northern Atlantic Index (TNA), Western Hemisphere warm pool (WHWP), and Southern oscillation index (SOI), indicated significant CIYL anomalies in some months. Furthermore, the regional yield simulation results using linear regression for the independent variables of year and various LOACI were satisfactory, with the average relative error ranging from 3.48% to 6.87%.

Knowledge of the responses of winter wheat yield to meteorological dryness/wetness variations is crucial for reducing yield losses in Henan province, China’s largest winter wheat production region, under the background of climate change. Data on climate, yield and atmospheric circulation indices were collected from 1987 to 2017, and monthly self-calibrating Palmer drought severity index (sc-PDSI) values were calculated during the winter wheat growing season. The main results were as follows: (1) Henan could be partitioned into four sub-regions, namely, western, central-western, central-northern and eastern regions, based on the evolution characteristics of the time series of winter wheat yield in 17 cities during the period of 1988–2017. Among them, winter wheat yield was high and stable in the central-northern and eastern regions, with a remarkable increasing trend (p < 0.05). (2) The sc-PDSI in February had significantly positive impacts on climate-driven winter wheat yield in the western and central-western regions (p < 0.05), while the sc-PDSI in December and the sc-PDSI in May had significantly negative impacts on climate-driven winter wheat yield in the central-northern and eastern regions, respectively (p < 0.05). (3) There were time-lag relationships between the sc-PDSI for a specific month and the atmospheric circulation indices in the four sub-regions. Furthermore, we constructed multifactorial models based on selected atmospheric circulation indices, and they had the ability to simulate the sc-PDSI for a specific month in the four sub-regions. These findings will provide scientific references for meteorological dryness/wetness monitoring and risk assessments of winter wheat production.