2.1. Principal component analysis of inter-site difference information
Through principal component analysis, the number of principal components is extracted and the principal components with characteristic value greater than 1 represent the sample information. The results showed that there are four principal components with different site information feature value greater than 1 (Figure 1), which can explain 81.61% of the total variance between sites (Table 3).In The first principal component, the coefficient of medium latitude, sunshine hours, average relative humidity and average rainfall is large, which can be summarized as meteorological factors.In the second main component, the coefficient of sowing period, tolerance weight and product temperature is large, mainly reflecting the variety factors.In the third main component, the effective product temperature weight is higher, and the coefficient of average precipitation and average relative humidity in the fourth principal component is higher (Table 4).It can also be found from the factor component diagram that the three factors of product temperature, latitude and sowing period have a high proportion in the principal component analysis (Figure 2).
2.2 Division of sorghum ecological zone
According to the results of principal component analysis, three factors of product temperature, latitude and sowing period were selected as the main factors of
intersite cluster analysis for cluster analysis, and the test points were divided into 5 regions (Figure 3)
Combined with the existing ecological zoning and planting habits, sorghum ecological zones are divided as follows: the main northern producing areas are divided into northeast, northwest and north China spring sowing zones, four ecological regions in northern spring and summer, and the south into one region, that is, 5 first-level ecological zones and 14 second-level ecological zones (Table 5).
2.3 Analysis of the yield performance of the participating varieties
Analysis of the average yield of 20 test varieties in 2020 and 2021 found significant differences between varieties in the two years.In 2020, the average yield of 20 varieties is 461.33kg/ mu, the coefficient of variation is 16.8%, and the change range is 267.92kg/ mu ~ 535.28kg/ mu, including the average yield of JN2 (535.28kg/ mu), LN3 (532.11kg/ mu) and LN11(521.99kg/ mu) is the highest, while the average yield of LZ18(267.92kg/ mu), LZ20(293.85kg/ mu) and HYZ (317.34kg/ mu) is the lowest(Figure4a).The average yield of 20 participating varieties in 2021 was 471.75kg/ mu, Slightly higher than that seen in 2020, The coefficient of variation was 16.11%, The change range is 260.80kg/ mu ~ 530.23kg/ mu, LN3(530.23kg/ mu), LN11(521.71kg/ mu), JNL5(520.49kg/ mu) three varieties ranked first, second and third respectively, Variars with lower yield in 2021 are consistent with 2020, they are three varieties: LZ (260.80kg/ mu), LZ20 (295.61kg/ mu) and HYZ (358.08kg/ mu).(Figure 4b)
There were also significant differences in yield between sites.The average yield of each test site in 2020 is 460.94kg/ mu, The coefficient of variation was 30.74%, The variation range is 203.96kg/ mu ~ 884.51kg/ mu, Except for Xinjiang Qitai (884.50kg/ mu) and Zhanjiang(830.31kg/ mu), The remaining sites of the three sites with the highest production are Kuerle (670.33kg/ mu), Zhangjiakou (658.83kg/ mu) and Chengde (657.37kg/ mu), The five sites with the lowest yield are Jiande (203.96kg/ mu), Nanning (214.56kg/ mu), Tongyu (261.22kg/ mu), Jinzhou (261.26kg/ mu) and Yancheng (286.3kg/ mu)(Figure5 a).In 2021, the average output of each is 471.76kg/ mu, the coefficient of variation is 25.12%, and the variation is 248.81kg/ mu ~ 695.21kg/ mu. The average sorghum yield of Tongliao (695.21kg/ mu), Kezuozhongqi(694.49kg/ mu) and Baicheng (673.46kg/ mu) is the highest, while the average yield of Yan'an (248.81kg/ mu), Yancheng (256.80kg/ mu) and Shijiazhuang (272.09kg/ mu) is low(Figure 5 b).
2.4 Advantages varieties in each ecological zone
The GGEBiplotGU package of R language was used to analyze the performance of different ecological zone species, and the double icon plot was fit to 86.67%.The "Which wins where" plot in the double icon pilot groups the suitable varieties according to the variety-environment relationship, and reveals the highest yield varieties within each group (Figure 6).
In North China and northeast China are Liaojie 3, Liaonuo11, JinNuoliang 5, Ji Wine No.2, Jinzhai 109, and Stanley 106. In terms of yield performance, the best variety is 3, followed by Liaonuo11; the best performance is Jinzhai 109, followed by Ji No.2. With comprehensive yield and stability, the best variety in North and Northeast China is 3, followed by Liaonuo11 (Figure 6).
Spring and summer sowing area, southern area suitable varieties 127, Jinnuo3, Liao 19. The output is Jinnuo3> 19> 127, and the stability is 127> Jinnuo3> 19. In general, Jinnuo3 is the most suitable variety in this region (Figure 6).
The suitable varieties of Fengza4 and good yield performance are Fengxia 4, Jicai 124, Xingliang No.2, Tongzi 108 and Liao37. The variety with the highest yield in this area is 4, the variety with the best stability is 108, and the best comprehensive performance is Liao37 (Figure 6).
2.5 Quality division of sorghum
2.5.1 Region performance of tannin content in sorghum
From the different test points, The 46 test points in 2020 ranged from 0.98% to 1.41%, The three test sites of AnYang (0.98%), DaTong (0.98%) and JinZhong (1.02%) had low tannin content, Pingliang (1.41%), Luoyang (1.39%) and Nanning (1.35%) have a high tannin content; The range of tannin content in the 38 test points in 2021 ranged from 0.86% to 1.59%, AoHan (0.86%), KuErLE (0.92%) and Datong (1.05%),Luoping (1.59%), Dongping (1.46%) and ZiBo (1.46%) had high tannin, Among them, the tannin in Datong in the past two years were relatively low.ANOVA showed that differences in tannin content reached significant levels both between cultivars and sites(Figure7).
According to the division of ecological zones and combined with the mean of tannin data in two years, the average tannin content of northeast spring sowing, northeast spring sowing, northwest spring sowing, northern spring and summer and southern regions were 1.64%, 1.44%, 1.49%, 1.54% and 1.64%, and the tannin content in North China was low(Figure 8)
2.5.2 Regional performance of the starch content in sorghum
In terms of location, the difference in starch content between regions is also significant. In 2020, starch content was 69.76% to 74.22%, and the variation range and variation coefficient (1.45%) were small, among which Zhejiang (69.76%), Nanning (70.23%) and Chaoyang (70.34%) were the smallest, while Zhangjiakou (74.22%), Chifeng (74.18%) and Xinzhou (73.52%) had the largest.In 2021, starch content in 38 sites was 69.13% to 79.83%, coefficient of variation was 2.18%, in Kuerle (69.13%), Shijiazhuang (70.11%) and Anyang (70.95%), and in Aohan (79.83%), Chengde (75.68%) and Qiqihar (75.66%)(Figure 9)
Starch content has obvious regional distribution, and the starch content gradually increases with the increase of latitude. The starch content in Inner Mongolia, Jilin and Heilongjiang is higher, with 73.38%, 73.16% and 72.89% respectively, while the starch content in Guangxi and Zhejiang is low.According to the division of ecological zones, the average starch content of the five ecological regions of Northeast China, North China, northern spring and summer and southern China is 72.73%, 73.12%, 71.82%, 71.88% and 71.81%, respectively, with North China and northeast spring sowing regions being high starch zones(Figure 8).
1.2.6.2 Performance of grain unit in sorghum
Grain unit indicates grain weight per unit volume, as indicated by g / l.From the point of view of each pilot, there are also significant differences in grain unit between the pilots. In 2020, there are 16 pilots with grain unit over 740g/l, and the top three are TaiAn (792.90g/l), TongLiao (791.83g/l) and JiNan (790.20g/l). the three test sites with the lowest grain unit are XingAnMeng (564.21g/l), ShuangYaShan (564.40g/l), NanNing (565.37g/l).In 2021, a total of 45 sites were calculated, and the range of capacity changes was 528.25g/l~803.05g/l, with KaShi (528.25g/l), GuYuan (619.11g/l) and XingAnMeng (666.69g/l) being lowest, while TongLiao (803.05g/l) , Baicheng (786.9g) and Gongzhuling (784.85g) having the highest capacity.In provinces, Guangdong, Gansu and Inner Mongolia had low capacity, 627.68g/l,693.39g/l and 704.52g/l, respectively, Over the 740g/l has ten provinces,including Hunan, Zhejiang, Yunnan, Jiangsu, Jilin, Sichuan, Heilongjiang, Guizhou, Hebei, and Ningxia(Figure 10).
According to the division of ecological zones, the average capacity weight of the five ecological areas of Northeast spring sowing, North China spring sowing, Northwest spring sowing, northern spring and summer and southern region is 742.65g,752.82g,709.15g,720.64g,741.87g respectively, among which the weight of North China spring sowing area and Northeast spring sowing area are the high grain unit area(Figure 8).