Characteristics and Spatial Inuencing Factors of Natural Regeneration within Aquaculture ponds in a Mangrove Forest

Regeneration is an important component of community succession and understanding regeneration 15 dynamics is essential for forest protection and recovery management. Mangroves are distributed 16 along coastlines and this unique habitat has resulted in very different regeneration process. This 17 study took Dongzhaigang mangrove forest in Hainan, China as the study area, considered the 10 18 years regeneration process in 8 abandoned aquaculture ponds, and the spatial factors that influence 19 the regeneration process are analyzed . The objectives were to: a) investigate the natural dynamics 20 of the mangrove regeneration process in abandoned ponds, b) determine the main spatial factors 21 affecting the natural regeneration process. The results showed that the number of species and 22 individuals showed a tendency to initially rise and then decline, with the maximum occurring at 6– 23 8 years. The results of a diversity index showed an initial rise, with stabilization then occurring 24 over a 8 year period. Aegiceras corniculatum and Sonneratia apetala were typical pioneer 25 mangrove species in the study area, while Bruguiera sexangula and Kandelia obovata were 26 representative species of late regeneration period. Spatial factors, including pond area and shape, 27 relative elevation, distance to a tidal creek and surrounding trees area played important roles in the 28 regeneration of mangrove in ponds. Finally, the study considered the current situation regarding 29 mangrove restoration in China and suggested that natural regeneration of mangroves is a good 30 management option.

individuals showed a tendency to initially rise and then decline, with the maximum occurring at 6-23 8 years. The results of a diversity index showed an initial rise, with stabilization then occurring 24 over a 8 year period. Aegiceras corniculatum and Sonneratia apetala were typical pioneer 25 mangrove species in the study area, while Bruguiera sexangula and Kandelia obovata were 26 representative species of late regeneration period. Spatial factors, including pond area and shape, 27 relative elevation, distance to a tidal creek and surrounding trees area played important roles in the 28 regeneration of mangrove in ponds. Finally, the study considered the current situation regarding 29 mangrove restoration in China and suggested that natural regeneration of mangroves is a good 30 management option. 31 Keywords 32

Spacial influencing factors 36
Community structure 37

Introduction. 39
With a distribution across estuary and coastal areas in the tropics and subtropics, mangrove forest 40 is critically important for the maintenance of biodiversity (AwuorOwuor et al., 2019). In recent 41 decades, mangrove forests have been seriously damaged worldwide (Menéndez et al., 2018) due 42 to aquaculture and urbanization, rising sea levels, changes in hydrologic processes, and so on. 43 Among which aquaculture ponds is considered to be the most serious threaten. Since 1980s, the 44 prawn industry has boomed in Southeast Asian countries, with more than 1.2 million hm 2 of 45 mangroves has been converted into fish ponds (Richards & Friess, 2016 (Bosire et al., 2008). 56 Natural regeneration is considered to be a better way for improving biodiversity in contrast to artificial 57 restoration(Levis et al., 2019), however less study has been conducted on the process of mangrove 58 natural restoration in abandoned ponds. Forestry studies showed that forest gaps are an important 59 feature of natural forests and are thought to contribute to species diversity by providing 60 opportunities for niche differentiation in modes of regeneration (Richards et al., 2011) . 61 Theoretically, when a gap formed, adjacent trees are able to quickly capture the vacant growing 62 space and close the gap (Yamamoto, 2000). The gap phase is considered to be the most important 63 stage of the natural forest regeneration process (Connell et al., 1985). The aquaculture ponds 64 within the mangrove forest are supposed to be the forest gap. When ponds were abandoned and 65 tidal process returned to normal, the surrounding mangrove plant propagators (including 66 propagules and seeds) will drift into the abandoned ponds along the tidal creek, and begin to settle, grow and develop into mangrove plant communities (Stevenson et al., 1999). Mangroves show 68 considerable regeneration ability in natural disturbances (storms, hurricanes), suggest 69 pioneer-phase characteristics (Alongi D M, 2008). The factors of soil and water characters that 70 influencing mangrove regeneration were discussed (Yanmei X., et al., 2021), however, very few 71 studies have focused on the natural regeneration of mangroves in pond rehabilitation and the 72 spatial factors that influence this process remain unclear. 73 In this study, we surveyed natural regeneration process within 8 ponds 10 years after being 74 abandoned in Dongzhai Harbor of Hainan, and spatial factors that affected the community 75 structures were analyzed. We aimed to tell : (I) How long does a regeneration process take from 76 abandoned ponds to stable mangroves community; (Ⅱ) How does mangrove composition change 77 during the first 10-year natural regeneration period in ponds of mangrove? (Ⅲ) How do ponds 78 spatial properties affect the composition of the regeneration community? 79

Study area 81
The Dongzhaigang mangrove wetlands reserve has the most continuous distribution of mangrove Since 1980's, around 160 ha mangroves were transformed into aquaculture ponds within the 91 reserve. While in recent decades, protection and restoration of mangrove forest has getting more 92 attentions, aquaculture in mangroves is forbidden and ponds within reserve were abandoned. To 93 accelerate vegetation restoration, artificial planting was the common way, while many adverse 94 impacts were noticed, such as limited biodiversity, introduced fast growing species, existing 95 habitat destroyed by large machines. Natural regeneration were more and more recommended by 96 scholars, while little of the natural regeneration process in this area was known. 97

Sample plots 98
Eight natural regeneration ponds which were abandoned in the year 2009 were selected for study 99 plot (see Fig.1 Ground diameter (D), height (H) and crown diameter (C) were measured in sample plots. 115 H ' (Shannon-wiener Index), which indicates distance-independent plant composition diversity, the 116 higher the Shannon Index, the higher the community biodiversity, and E(Pielou index), which 117 indicates the spatial distribution uniformity of plants, the greater the Pielou Index is, the more 118 uniform the distribution of individual number is and the higher the biodiversity is. In this study, 119 these two indices were used to analyze the structure of communities (Lei & Tang, 2002 ). 120

Natural regeneration rate 149
The results showed that after 10 years natural regeneration, 74.88% of ponds returned to 150 mangroves forest on average, among which B pond regenerated 97.15% and F pond 37.25%, 151 which was the maximum and minimum. The rate of mangroves increasing within 8 ponds could 152 be seen in figure 4. 153 By comparing areas in different years, it showed that the average increasing rate was 6.89% per 154 year, while except the pond E, which increased quickly in 2019, the most rapid increasing of 155 mangroves happened during the year of 2013 to 2015, the time about 4-6 years after ponds being 156 abandoned (see Fig. 4), . 157 Hierarchal Clustering Analyses (see Fig. 5) showed that the regeneration rate between A and H 158 pond has the greatest similarity, followed by C and E pond, while similarity between F pond and 159 other ponds is the least. 160

Community composition and structure 161
It was found that the number of species increased over the first several years, reaching the average 162 Rhizophora styolsa and Kandelia obovata increased respectively from 8.4% to 33.3% and 3.0% to 181 8.7% during the same periods, meaning that with a change of environment, these species began to 182 adapt to the changes. The introduced species, Sonneratia apetala, experienced an increasing from 183 6.8% in 2011to 14.4% in 2015, and followed by a decreasing to 9.5% in 2019, indicating that 184 Sonneratia apetala is a pioneer species in the mangrove regeneration processing, and with the 185 stability of the community structure, its importance in the community gradually decreased. 186 Species such as Lumnitzera racemosa and Acanthus ilicifolius appeared only in 2015, meaning 187 that the seeds of these species can reach ponds area, but are less competitive than the other 188 species. 189

Influence of important spatial factors 190
Distance to tidal creeks (see Tab

Discussion 199
The regeneration rate after 10 years of natural restoration showed that more than 70% of the 200 cultivation ponds have been restored to mangroves(see Fig.4). That means mangrove community 201 can be regenerated by natural processing in the aquaculture ponds that located in mangrove forest. 202 During the 10 years of natural regeneration of mangroves, a total of 10 species of mangrove plants 203 appeared in the cultivation ponds (see Fig.7). Among them, there were 6 common mangrove 204 plants, including Aegiceras corniculatum, Rhizophora stylosa, Bruguiera gymnorrhiza, Bruguiera 205 sexangula, Kenaelia obovata and Avicennia marina, which are consistent with the surrounding 206 mangrove communities. What is noteworthy is that the max percentage of Sonneratia apetala in 207 the regeneration process was 15%, but there were no surrounding parent trees of S. apetala, 208 meaning that it has strong diffusion ability. 209 The mean of Shannon diversity and Pielou index of the 8 ponds were 1.5 and 1.9, and kept stable 210 since 2017(see Fig.6). It mean that the stable and original community structure can be restored by In this study, we found that natural regeneration 218 occurred in all ponds, and the species diversity was very high and community composition was 219 very similar to that of primary forest. So, within aquaculture ponds in mangrove area, natural 220 regeneration is highly recommended in future mangrove restoration projects. 221 The cluster analysis of the regeneration process within 8 ponds showed that the ponds with similar 222 area and shapes had the more similar regeneration process (see Fig.5), and smaller area and 223 area/perimeter were more conducive to natural restoration in ponds((see Tab.1), which indicated 224 that the natural regeneration process of the mangrove forests was affected by the area and shape of 225 the ponds. The area and shape of different ponds can affect seed germination and seedling 226 settlement through different amount of seed interception, water exchange frequency and soil 227 flooding time (Mcnab et al., 2020;Saraiva et al., 2020), so as to affect the natural regeneration 228 process of aquaculture ponds. 229 Mangrove habitats are within coastal zones, and seed dispersal is largely due to transport by tidal 230 creeks. The results showed that the distance to tidal creeks were important for mangrove 231 community structure (see Tab.1). Seed dispersal is at the core of the Janzen Connell hypothesis 232 that explains the high diversity of tree species in tropical forests (Fox et al., 1976;Hyatt et al., 233 2003). Water are considered to be important media for the transmission of seeds (Andresen et al., 234 2005; Rotundo et al., 2005). Species such as A. corniculatum, with a huge floating seed capacity, 235 have an absolute advantage during first years (see Fig. 7). Its seeds are carried by tidal water and 236 reach open spaces, becoming retained in ponds, while those species with seeds that have less 237 ability to float, such as B. sexangula, K. obovata, and Rhizophora stylosa, would have fewer 238 opportunities to occupy ponds at the very beginning of the regeneration process (see Fig. 7). 239 The relative elevation and adult tree area have a significant effect on community structure during 240 the regeneration process (see Table 1). Within-gap heterogeneity should be taken into account 241 when considering the factors that affect natural regeneration (Gray et al., 1996). Heterogeneity 242 may be caused by many factors (Chen et al., 2005), such as micro-topography, soil 243 physical-chemical characteristics, and shelter by the surrounding forest. In a mangrove forest, the 244 ponds with the greatest difference in topographical elevation (height difference) will trap most 245 mangrove seeds, and therefore relative elevation had the greatest influence on community 246 composition at the very beginning of the regeneration process (see Fig.6). As the plants grow, the 247 ground flattens out, and the effect of relative elevation gradually decreases. 248

Conclusion 249
By 10 years monitoring, we concluded that it would take about 6-8 years to generate naturally 250 from an abandoned aquaculture pond to a stable mangrove plant community. Although compared 251 to artificial forest, the natural regeneration process requires longer time, less disturbance and 252 richer biodiversity are expected. 253 Aegiceras corniculatum is an important species in the whole regeneration processing. It is not only 254 the main constituent species, but also the most important pioneer species, played an important role 255 in early settlement and habitat modification. While mangrove species, such as Rhizophora styolsa 256 and Kandelia obovata, with less propagules, are highly competitive and therefore occupy an important proportion in the final stable communities. The introduced species, Sonneratia apetala, 258 is also a pioneer species, the invasion risks should be taken into consideration in future pond 259

rehabilitation. 260
There are many factors that influence the regeneration process, which should be taken into 261 account. Nevertheless, this study demonstrated the importance of tidal creeks around ponds. A 262 perfect tidal creek system is indispensable for successful natural regeneration within ponds. 263

Acknowledgment 264
The study was supported by Grant No.  Increasing rate of mangroves area within each pond in the study site  Percentage of mangrove species composition in different regeneration years