The Changes of Soil Seed Banks In Time Series Before And After The Invasive Artemisia Trida Was Removed From Grassland

10 Abstract Artemisia trifida (giant ragweed) is an invasive weed with an expanding distribution area. 11 In recent years it has been found to invade grasslands, bringing great challenges for effective weed 12 control and restoration of native herbage. Although it has been reported that plant invasion can cause 13 a decline in species richness and biodiversity in native seed banks, which may eventually lead to the 14 depletion of native seed banks, few location- and species-specific case studies have been conducted 15 regarding the dynamic characteristics of the invaded seed banks from invasion back to restoration. 16 The purpose of this study was to compare and quantify the seed banks of grassland communities after 17 (1) giant ragweed invasion for 0-8 years, and (2) giant ragweed removal, in Yili Valley, Xinjiang, 18 China. The results showed that the duration of invasion determined whether giant ragweed could pose 19 a significant threat to the native community seed bank. The seed bank density of native community 20 had significantly decreased by 30.44% after 4 years of invasion, and in the sixth year, the species 21 richness in the seed bank had decreased significantly by 12.36%. After the invasion had lasted for 22 eight years, the seed bank density of the native community had decreased by 83.28%, the species 23 richness in the seed bank decreased by 39.33%, and the seed bank tended to be homogeneous. After 24 the giant ragweed was removed, the potential for restoration was limited by the residual seed bank. 25 Three years after the restoration, although the density of seed banks increased significantly, new 26 growth was dominated by weedy species, rather than crucial components of the grassland habitat. This study is of great significance to the control of giant ragweed and the restoration of grassland 28 vegetation invaded by giant ragweed.


59
Non-native invasive plant species are now recognized a major threat to ecosystems worldwide 60 (Ehrenfeld 2010

67
The germination of mature seeds may be delayed for an indefinite period. During this time, the 68 seeds on the surface of or within the soil are likely to form a reserve, known as a soil seed bank 69 (Fenner and Thompson 2005). The seeds of these reserves can be used for risk diversification, such as   87 Moreover, research also shows that the resistance of the communities that have been invaded over 88 long time periods is reduced due to the decline in diversity, so, these communities are easily invaded  4 mainly invades wasteland, roadside, and riverside habitats, as well as important production farmlands 103 (Gudžinskas 1993;Follak et al. 2013;Chauvel et al. 2015

116
In order to solve this problem, in this study, we compared the composition of seed banks (1) with 117 different invasion years (8 consecutive years) and (2) at different times after the control (three

133
Leguminosae, Apiaceae, and Asteraceae species distributed among them (see Table 1 for details). We

137
Because of the rapid diffusion ability of giant ragweed, it formed many populations with different 138 invasion years, and some sections were selected as long-term fixed plots for the experimental study. A total of 480 soil samples were collected from seven sites that had been invaded by giant 147 ragweed for 2-8 years, and the data that invaded by giant ragweed for 1 year were obtained around the 148 site that invaded for 2 years, and also include the data of the native non-invaded areas. Each site 149 contained multiple "patch communities" formed by the invasion of giant ragweed, and that 3 patch 150 communities were tested at each site. Referring to the sampling methods of Robertson (2012) and 151 Ferreras (2015), we randomly set up four 5×5 m plots in each invasive community (in the middle of 152 the invasive community, the edge area was avoided, because the edge may be newly established). The 153 relative coverage of giant ragweed was estimated at the growth peak of mid-August. In the first ten 154 days of November after the end of the growing season, four sampling points were selected from each 155 plot, measuring 20 × 20 cm, 5 cm deep (referring to Dong (2020), as our previous research showed 156 that most of the seeds of giant ragweed gathered at 0-5 cm depth, and Robertson (2012), Dairel (2020) 157 referred that the seeds of grassland herbs were mostly on the surface of soil). Soil was collected, and 158 the four soil sampling points of each plot were mixed separately. The seed bank data of the native 159 non-invaded areas were obtained from around the sites that had been invaded for 1, 3 and 7 years (the 160 data obtained from the site of one year of invasion can be directly used here, because it had just 161 settled here and had not caused harm to the native species and had not spread more seeds into the 162 soil.), where three 20×20 m plots were set in the noninvaded areas and four 5×5 m quadrats were 163 randomly set in each plot. The sampling method was the same as above.

164
For identification and quantification of soil seeds before the collection of soil seed bank samples, 165 all types of spermatophytes on the ground were investigated and recorded, and some seeds were 166 collected during the seed setting period to compare and identify the species in the seed banks. After

214
The seed bank density in noninvaded area was between 21500 and 37800 seeds per m 2 (F=2.775, 8 reaching 83.75% (±79%) in the fourth year, reaching 100% in the sixth year of invasion (Fig. 3a). The 218 seed bank density of the native species community decreased with the increase of the invasion 219 duration ( Fig. 3b; R 2 =0.9336, p<0.01). There was no significant difference between the seed bank 220 density of the community and that noninvaded area in the first three years of invasion, and the seed 221 bank density decreased significantly by 30.44% from the fourth year (Fig. 3b). However, the seed 222 bank density of the dominant species (Festuca elata and Lolium perenne) decreased significantly by 223 16.75% from the third year (Fig. 3d). The species richness in the seed banks began to decrease 224 significantly by 12.36% in the sixth year, and then decreased year by year (Fig. 3c). In terms of 225 species composition, the seed bank density of Gramineae and non-Gramineae decreased continuously 226 with the increase of invasion duration (Fig. 3e); The seed bank density of species with different life 227 cycles (perennial, biennial and annual) also showed a continuous downward trend (Fig. 3f).  (Table 2).

247
After chemical control, the relative seed bank density of giant ragweed decreased rapidly until it 248 disappeared, accompanied by the rapid increase of seed bank density of native plants (Fig. 4a). In the 249 third year of restoration, the density of seed bank had reached the same level as that of noninvaded 250 area (Fig. 4a) and the coverage of native community had also been restored (Fig. 4c). However, the 251 families that had disappeared in the duration of invasion could not be reestablished, and the species 252 richness decreased by 41.57% compared with that before invasion (Fig. 4b).

297
We compared pre-and post-invasion data to examine the cause-effect nature of invasion. Our 298 investigation of non-invaded areas revealed that 20-25 species in the original non-invaded 299 communities, and the seed bank density in November was between and 21500-37800 seeds per m 2 . As 300 seen in Figure[

303
These results indicate that the duration of invasion is key to reducing the native species diversity

316
There are special requirements for the chemical control of giant ragweed invading grassland. Our 317 survey found that the species richness of grassland was higher than that of farmland, roadside, 318 wasteland, and other habitats, and as animal husbandry land, the requirements for safe drug use are 319 more stringent. Therefore, it is necessary to develop special chemical control methods based on their 320 growth and reproduction characteristics.
produced every year, which is an important reason why giant ragweed can cause damage to the 324 grassland. With the increase in invasion years, a dense forest population was formed in the 325 aboveground portion of the studied ecosystem. The relative coverage of the population expanded 326 explosively in the fourth year, reaching 83.75%, and reaching 100% by the sixth year (Fig. 3a). The 327 plant height of giant ragweed invading grassland is generally 2-4 m, which is higher than that of 328 native plants (Table 1). Therefore, giant ragweed has more advantages in the competition for light  (Table 2), so as to achieve effective weed control.

352
The first goal in prevention and control is to remove invasive plants from the community.

353
Second is to ensure that the community can recover species diversity before re-invasion (Klaus et al.

357
The results showed that the original species diversity could not be restored in a short time after  re-established, and proliferated in large numbers in the community, occupying the blank patch. In the 12 higher than that of non-invaded area after 3 years of restoration in weed control area (Table 3)  Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.

Figure 2
The distribution characteristics of (a) seed bank density and (b) species richness in noninvaded area are represented by box chart Figure 3 Effects of giant ragweed invasion duration on aboveground relative coverage and underground seed bank of native community. (a) the changes of relative coverage of giant ragweed with invasion time; the changes of (b) seed bank density and (c) species richness of native communities; and the changes of seed bank density with invasion time of (d) two dominant species and (e) gramineae/non-gramineae (f) and that species with different life cycles. Different letters indicate signi cant differences (p<0.05) using a least signi cant difference test Figure 4 The restoration of (a) native seed bank density, (b) species richness in the seed bank and (c) native community coverage after removal of giant ragweed. Different acronyms represent different restoration times that NA (Noninvaded area), BR (Before restoration), FR (The rst year of restoration), SR (The second year of restoration), TR (The third year of restoration). Different letters indicate signi cant differences (p<0.05) using a least signi cant difference test