Abstract: The difference of interdecadal variations, coefficient of variation and tendency ratio between the observed yield of winter wheat from 123 agrometeorological observation stations and the statistical yield of winter wheat at county level where the observation station was located from 1991 to 2017. The proportion of average winter wheat planting area in each county in five years(2006−2010) was used as the weight factor to integrate the observed yield and statistical yield at province level, at the same time using the announced yield at province level from National Bureau of Statistics. The interdecadal variations and tendency ration of three different yields at provincial level were compared and analyzed. The results showed that:(1) the number of high yield counties increased significantly, and low yield counties decreased significantly in both observed yield and statistical yield counties. The two yield were both high yield years in the 2010s, and the difference between the two reached peak value in the 2000s. (2) The coefficient of variation of observed yield at the county scale was higher than the statistical yield. The coefficient of variation of statistical yield in 49 counties were less than 0.20 and only 8 were greater than 0.40, while 72 statistical yield counties were less than 0.20 and only 9 were greater than 0.30. The coefficient of variation of statistical yield in all counties in Xinjiang and Shandong provinces were less than 0.30. (3) The tendency ratio of 73 observed yield counties showed a significant increase mostly concentrated in the major producing provinces such as Hebei, Henan, Shandong, Jiangsu, and Anhui; and 100 statistical yield counties showed the same significant increase. The tendency ratio of observed and statistical yield in 72 counties passed the significance test at the same time. (4) The 2000s were the high yield years for both observed and statistical yield at provinces level and 1990s were the low yield years. The average of the observed yield in every 10 years was higher than the average of the statistical yield in Shandong, Anhui, Hebei, Jiangsu, Shaanxi and Shanxi province. (5) Eight provinces had passed the significant test on tendency ration of the observed yield at the provincial level except for Xinjiang and Shanxi province. While tendency ration of the statistical yield and the announced yield in all provinces had passed the significant test and the yield growth was positive. In general, the winter wheat yield series based on the observed yield could provide a new data source for yield forecast.

Key words: Winter wheat, Agrometeorological observation stations, Observed yield, Statistical yield , Area weight factor integration