Gap edge canopy buffering of throughfall deposition in a subalpine natural forest

Siyi Tn, Qing Dong, Xingyin Ni, Ki Yue, Shu Lio, Fuzhong Wu,＊

a Key Laboratory for Humid Subtropical Eco-Geographical Processes of the Ministry of Education, School of Geographical Sciences, Fujian Normal University, Fuzhou,350007, China

b Ecological Security and Protection Key Laboratory of Sichuan Province, Mianyang Teachers' College, Mianyang, 621000, China

Keywords:Canopy exchange Dry deposition Throughfall Subalpine forest Precipitation deposition

ABSTRACT Base cation loads are rarely considered in forest gap edge canopies,but they can play critical roles in capturing or buffering atmospheric deposition in forests with frequent gap disturbances,such as subalpine forests.We selected an expanded gap edge canopy and a closed canopy in a subalpine natural forest on the eastern Tibetan Plateau.The throughfall deposition and canopy exchange processes of common base cations (K+, Ca2+, Na+, and Mg2+)were continuously studied over two years. The results showed that the enrichment ratio and fluxes had lower levels of base cations in the gap-edge canopy than in the closed canopy, which indicated that base cations were concentrated more in the closed canopy than in the gap-edge canopy. Although Ca2+ in the gap-edge canopy showed a higher net throughfall flux,the annual net throughfall fluxes of K+,Na+and Mg2+within the gap-edge canopy were 1.83, 6.75 and 2.95 times lower than those in the closed canopy, respectively. Moreover, dry deposition fluxes of base cations significantly decreased in the gap edge canopy compared to those in the closed canopy,and the decreasing tendency was more significant during the snowy season than during the rainy season.Overall, these results suggest that the amount of base cations in subalpine natural forest ecosystems may be overestimated when the throughfall deposition of ions in gap edge canopies is ignored.

1. Introduction

The load of base cations(including K+,Ca2+,Na+and Mg2+)can be transferred via atmospheric (wet/dry) deposition in forests (Tan et al.,2018) and represents an important nutrient flow into the forest floor while buffering acidification and nutrient imbalance in soil (Chadwick et al., 1999). Precipitation washes the deposits accumulated on the canopy or exchanges with leaves, which changes the base cations in precipitation to the forest soil (Parker, 1983; Shen et al., 2013).Currently, increasingly many forest gaps are created by fallen trees,anthropogenic disturbances and extreme climate disasters (such as storms and snow damage) during forest regeneration (McCarthy, 2001;Kajimoto et al., 2002). The gap edges can be considered “rings” of the ecotone,and environmental change also occurs at edges,which supplies a different and much greater microenvironmental heterogeneity and throughfall fluxes compared to the closed canopy(Lindberg and Owens,1992;Scharenbroch and Bockheim,2007).These changes will affect the nutrient cycle in subsequent hydrological processes in the forests,but the base cations from throughfall deposition in forest gap edge canopies are currently poorly understood compared with that in closed canopies.

Generally,the net throughfall flux of base cations is highly correlated with dry deposition and canopy exchange processes, and it can be estimated by a widely adopted canopy budget model in various forest ecosystems (Ulrich, 1983; Balestrini et al., 2007; Van Langenhove et al.,2020). On one hand, the inherent features of canopies, such as canopy cover,may affect the variability in dry deposition and canopy exchange processes (Scheer, 2011). Specifically, higher canopy cover provides a large effective exchangeable canopy area and results in more dry deposition(Fenn and Bytnerowicz,1997;Mukesh et al.,2017).On the other hand,the dual effects on the physiological function of K+,Ca2+,Na+and Mg2+and seasonal canopy phenology alter the canopy exchange and dry deposition process. The canopy often performs with high physiological activity, and rainfall can leach a large amount of mobile soluble substances such as K+and Mg2+from tissues during the growing season(Polkowska et al.,2005;Scheer,2011;Sun et al.,2021).Conversely,the defoliation and dormancy of vegetation limits the canopy exchange of base cations in winter, which results in less leaching than in the rainy season (Zeng et al., 2005). Furthermore, the enrichment levels of these base cations of observed ions widely vary in different forest ecosystems(Parker,1983;Shen et al.,2013;Moslehi et al.,2019;Sun et al.,2021).

Previous studies have confirmed that base cations from throughfall deposition are definitively different due to slight changes in site characteristics (such as aspect, slope and altitude) (Holder, 2004; Siegert et al., 2017; Sun et al., 2021). Sites at higher altitudes have greater dry deposition flux and can produce larger total volume throughfall flux through condensation of cloud water(Holder,2004),which can enhance the base cation flow. In addition, Sun et al. (2021) found that the dry deposition fluxes increased in the upper slope, while canopy exchange increased in the lower slope, so the net throughfall flux of base cations varies between the two slopes.Although the pathways of water in the gap edge canopy and closed canopy are similar(Van Stan et al.,2020),both canopies represent different canopy features and site characteristics.The gap edges are sparsely covered with shrubbery, and trees may show a small magnitude of base cation fluxes relative to closed canopies across seasons. However, the lack of attention to gap edges creates a potential risk in overestimating the base cation fluxes into forest floors, which limits our understanding of nutrient cycling in hydrological processes.

As an important part of the freshwater reservation zone in the Yangtze River Basin,the subalpine forests in southwestern China play a major role in regulating the regional climate,maintaining natural biodiversity,etc.(Yang et al., 2006). Our group has found that increasingly many forest gaps are formed by high-frequency climate disturbances such as wind or snow damage in subalpine forests(Wu et al.,2013).Our previous study also demonstrated that base cations in throughfall were one of the main sources of nutrients in subalpine forests (Tan et al., 2018), while information on how the gap edge canopies affect the input of base ions via throughfall in subalpine forests with seasonal changes is neglected but crucial.Thus,we estimated the throughfall deposition of base cations for a gap edge canopy and a closed canopy over two years.We hypothesized that the gap edge canopy could display a lower enrichment ratio and lower fluxes of K+, Ca2+, Na+and Mg2+than the closed canopy. Specifically,we aimed to 1)quantify and compare the differences in fluxes of K+,Ca2+,Na+and Mg2+in throughfall between forest edge canopy and closed canopy and 2) analyze the seasonality (rainy and snowy season)dynamics on the base cation deposition between the two canopies. In conjunction with field observations, the results can facilitate a more complete picture of ecohydrological cycles in forest ecosystems from the perspective of gap edge canopies.

2. Materials and methods

2.1. Site description

This research was conducted at the Long-term Research Station of Subalpine Forest Ecosystems, Bipenggou Nature Reserve(102°53′–102°57′E,31°10′–31°19′N;2,458–4,619 m a.s.l.),Li County,Sichuan,southwestern China.According to the previous meteorological data of this area, the mean air temperature is 2–4°C, and the annual precipitation in this area is 801–850 mm.The dominant coniferous tree species are Abies faxoniana, Sabina saltuaria, Betula albosinensis and Larix mastersiana. The shrub layer comprises Salix cupularis, Salix paraplesia,Rosa omeiensis,and Berberis silva-taroucana in this area(Fu et al.,2017).

Based on previous field investigations by Wu et al. (2013), the expanded gaps(the canopy gap and the area that extends to the bases of around canopy trees)cover 12.60%of the sampling sites,and the shape of the forest gap is elliptical.We established three similar replicated plots of 1 ha(100 m×100 m)with homogeneous locations,slopes and aspects in L. mastersiana-A. faxoniana forest stands and a distance of 500 m between each plot.Moreover,the extended gap canopy and closed canopy were deployed at each plot to observe precipitation and throughfall,and the closed canopy was approximately 20 m from the forest extended gap edge canopy.Only one open field(20 m×20 m)without tree and shrub layers 50 m from the L. mastersiana-A. faxoniana forest was selected tomonitor precipitation. Table 1 displays the detailed canopy characteristics of the studied plots.

Table 1 Site characteristics of subalpine natural forests in the gap edge canopy and closed canopy.

2.2. Water sample collection

Rainfall and throughfall during the rainy season:The rectangular PVC drains as a collector (the collection area was 400 cm × 16 cm for each surface). Each collector was installed at a height of 1 m from the forest floor, and the collector was connected to a 15-L polyethylene (PE)bucket. Five collectors were arranged in the open field, forest gap edge canopy and closed canopy of each plot(Fig.1).

Snowfall and throughfall during the snowy season:The cone-shaped collector devices (top diameter: 1 m; bottom diameter: approximately 0.2 m) comprised PVC and gridding cloth; moreover, the collector was connected to a 20-L polyethylene(PE)bucket at the bottom.During the snowy season, the five collectors were deployed in the same manner as during the rainy season(Fig.1).

The experiment lasted two years from August 2015 to July 2017,and rainfall,snowfall and throughfall were collected.In the first year(August 2015 to August 2016), we collected water samples from each rainfall event during the rainy season and once a month during the snowy season(due to the harsh natural conditions during snowfall).The snowy season lasted six months from November 2015 to April 2016. A graduated cylinder measurement of water volume was performed for the first year.During the second year(September 2016 to July 2017),the snowy season lasted five months from October 2016 to February 2017. A micrometeorological station (HOBO U30) was established to automatically detect the volume of water; then, water samples were collected once a month.All samples were collected with 250-mL polyethylene plastic bottles(washed with deionized distilled water).After collection,the impurities in the collectors were rinsed by washing with distilled water.

2.3. Chemical analysis and calculations

All water samples were placed in the laboratory, stored in a refrigerator at a temperature of 4°C after every collection, and filtered by a filter membrane (with 0.45 μm). The pH of the filtered sample was adjusted to 1–2 afterwards. The concentrations of K+, Ca2+, Na+and Mg2+were measured by an atomic adsorption spectrophotometer(AA-7000, SHIMADZU, Japan). Ulrich (1983) and Draaijers and Erisman(1995)proposed a canopy budget model that uses a tracer ion-Na+to evaluate the ion exchange process that occurs in a forest canopy.This model assumes that Na+was ignored in dry deposition and canopy exchange input;in addition,the deposition efficiency of Na+is equivalent to that of the investigated ions.

Canopy budget model:

Fig. 1. Design model of the precipitation experiment in a subalpine natural forest.

where TFD-throughfall deposition; PD-precipitation deposition;DD-dry deposition;NTF-net throughfall flux;and CE-canopy exchange(value for negative/positive present uptake/leaching). When the net throughfall deposition flux of Na+is negative,the dry deposition value is zero.

The enrichment ratio was calculated by dividing the concentration of base cations in the throughfall by the precipitation.

2.4. Statistical analysis

Statistical analysis was performed using IBM SPSS statistical 20.0 software (IBM SPSS statistics Inc., Chicago, Illinois, USA). One-way analysis of variance (ANOVA) was used to test the concentration and fluxes among the bulk precipitation and throughfall in the forest gap edge canopy and closed canopy.Univariate ANOVA testing was performed to study the effect of the season and canopy on the concentration,fluxes,net throughfall flux, dry deposition flux and canopy exchange flux of base cations. The concentration, fluxes, and net throughfall flux of the two seasons (rainy and snowy seasons)were compared with independent Ttests.All figures were generated in Version 4.1.0(R Core Team,2021).A significance level of P ＜0.05 was set.

3. Results

3.1. Concentrations of base cations K+, Ca2+, Na+ and Mg2+ in the throughfall

The monthly variation in the mean concentration of base cations is shown in Fig. 2, which shows that the base cation concentrations were enriched in throughfall in the gap-edge canopy and closed canopy compared with those in bulk precipitation. Meanwhile, the enrichment ratio was 1.27–2.51 in the gap-edge canopy and 1.63–3.20 in the closed canopy(Fig.3b).In terms of the canopy,the mean concentrations of base cations in the gap-edge canopy were significantly lower than those in the closed canopy during the two years of observation(Fig.3a).Canopy had a notable influence on the concentrations of base cations(except for K+,Table 2).Seasonally,the concentrations of K+,Ca2+and Mg2+were more than 1.1 times higher during the rainy season than during the snowy season under both canopies(Table 3).

Fig. 2. Monthly variation in the mean concentrations of K+ (a), Ca2+ (b), Na+ (c) and Mg2+ (d) in bulk precipitation (BP) and throughfall in the forest gap edge canopy (GE) and closed canopy (CC) during the two-year observation period from August 2015 to July 2017. The error bars represent the standard deviation of the mean (n = 5).

Fig. 3. Mean concentration of base cations in bulk precipitation (BP) and throughfall in the forest gap edge (GE)canopy and closed canopy (CC) during the two-year observation period from August 2015 to July 2017 (a). Different lowercase letters denote significant(P ＜0.05) differences in the concentrations of base cations in BP, GE and CC (a).Enrichment factor of base cations in the forest gap edge (GE) canopy and closed canopy (CC) (b). Different lowercase letters denote significant (P ＜ 0.05)differences in the concentrations of base cations in GE and CC(b). The error bars represent the standard deviation of the mean (n = 5).

Table 2 Univariate ANOVA results to study the effect of canopy and season on the concentration,fluxes,and net throughfall flux of base cations.Bold P values are significant(P＜0.05).

Table 3 Mean concentration of base cations from bulk precipitation(BP)and throughfall in the gap edge (GE)canopy and closed canopy(CC).

3.2. Fluxes of K+, Ca2+, Na+ and Mg2+ in throughfall

After passing through the canopy, the annual throughfall fluxes of base cations in the gap-edge canopy and closed canopy were higher than those in bulk precipitation(except for Ca2+,Fig.4).Moreover,the annual fluxes of K+, Ca2+, Na+and Mg2+in the gap-edge canopy were lower than those in the closed canopy. Seasonally, the fluxes of base cations were greater during the rainy season than during the snowy season for both canopies. The interaction between canopy and season had an insignificant effect on the base cation fluxes(Table 2).

3.3. Net throughfall fluxes of K+, Ca2+, Na+ and Mg2+

The annual dry deposition fluxes of K+,Ca2+and Mg2+significantly(P ＜0.05) increased from the gap edge canopy to the closed canopy,especially during the rainy season (Fig. 6a, b and c). However, the interaction between canopy and season had little significant effect on the dry deposition fluxes of K+,Ca2+,Na+and Mg2+(Table 1).Additionally,the annual canopy exchange fluxes were lower in the gap-edge canopy than in the closed canopy. Seasonally, the canopy exchange of K+and Ca2+in the rainy season was lower in the gap edge canopy than in the closed canopy(Fig.5d and e).

Fig. 4. Fluxes of K+ (a), Ca2+ (b), Na+(c), and Mg2+ (d) in bulk precipitation(BP), throughfall in the forest gap edge(GE) canopy and closed canopy (CC)during the annual rainy season and snowy season. The lowercase letters indicate significant differences in fluxes of base cations in BP, GE and CC (P ＜0.05). The capital letters indicate significant differences in fluxes of base cations between different seasons in the same canopy (P ＜0.05). The error bars represent the standard deviation of the mean (n = 5).

Fig.5. Dry deposition(DD)fluxes of K+(a),Ca2+(b),and Mg2+(c)and canopy exchange(CE)fluxes of K+(d),Ca2+(e)and Mg2+(f)in throughfall in the forest gap edge (GE) canopy and closed canopy (CC) during the annual rainy season and snowy season (CE flux value for negative/positive present uptake/leaching). The lowercase letters indicate significant differences in DD and CE of K+,Ca2+and Mg2+in the GE and CC(P ＜0.05).The capital letters indicate significant differences in DD and CE of K+, Ca2+ and Mg2+ between different seasons in the same canopy (P ＜0.05). The error bars represent the standard deviation of the mean (n = 5).

The net throughfall flux of base cations(except for Ca2+,Table 1)was significantly influenced by the canopy.The annual net throughfall fluxes of K+, Na+and Mg2+in the gap-edge canopy were 1.83, 6.75 and 2.95 times lower than those in the closed canopy, respectively (Fig. 6).Moreover, higher fluxes of base cations were observed during the rainy season in both canopies (Fig. 6). The interaction between canopy and season did not exhibit obvious effects on the net throughfall flux of base cations(Table 1).

Fig. 6. Net throughfall flux (NTF) of K+(a), Ca2+ (b), Na2+ (c) and Mg2+ (d) in throughfall in the forest gap edge (GE)canopy and closed canopy (CC) during the annual rainy season and snowy season. The lowercase letters indicate significant differences in the NTF of base cations between forest gap edge canopy and closed canopy (P ＜0.05). The capital letters indicate significant differences in the NTF of base cations between different seasons in the same canopy (P＜0.05). The error bars represent the standard deviation of the mean (n = 5).

4. Discussion

The findings here are consistent with our hypothesis, which demonstrates that the gap edge canopy can display a lower enrichment ratio of base cations than the closed canopy(Fig.3).The base cation fluxes in the gap-edge canopy were slightly lower (no significant difference) than those in the closed canopy(Fig.4).Seasonally,the fluxes of base cations were higher during the rainy season than during the snowy season under both canopies.Furthermore,such estimations of base cation fluxes in gap edge canopies help accurately quantify the amount of nutrients in forest floors from hydrologic pathways.

Numerous studies have reported that the ions in throughfall can be greatly enhanced when precipitation passes through the canopy (Levia and Frost, 2006; Berger et al., 2008; Tonello et al., 2021). We also observed that the mean concentrations of K+, Na+and Mg2+in the throughfall of both canopies were 1-3-fold greater than those in precipitation (Fig. 3). In addition, the canopy characteristics of the closed canopy and gap-edge canopy enhanced the differences in base cation concentrations and fluxes. Gautam et al. (2017) reported that a higher canopy surface could provide a large area to interact with precipitation.A multilayer spatial structure can lead to overlapping canopies and additional opportunities for particle interception (Siegert et al., 2016). As a result, precipitation passes through the closed canopy with higher canopy closure,whereas more ions can be washed and leached into the soil than through a gap-edge canopy.Moreover,water migration is one of the important material transport processes in forests. The soluble ions that flow from throughfall may be the main and most convenient pathway of nutrient circulation in the linkage between forest and soil and represent an essential supplement to available soil nutrient pools (Parker, 1983;Kalbitz et al.,2000; Liu et al., 2021).

A large amount of dry or wet deposition captured by the forest canopy plays important roles in the forest nutrient cycle(Oda et al.,2009,2019).Based on the canopy budget model, the dry deposition and canopy exchange (leaching/uptake) process can be distinguished, which is essential in further understanding the discrepancies in net nutrient flux in throughfall as affected by canopies. Since Na+is usually transported by atmospheric marine salts (L¨ovblad et al., 2000), Na+deposition can be negligible in areas with low marine deposition levels. Therefore, the input of Na+from throughfall is close to that in bulk precipitation (Sun et al.,2021).To date,the canopy budget model has extensive application in various forest ecosystems (Fan and Wei, 2001; Adriaenssens et al.,2012; Tonello et al., 2021). We also observed insignificant differences between the enrichment degree of Na+fluxes in throughfall (4.32 kg·ha-1in the gap-edge canopy and 4.65 kg·ha-1in the closed canopy,Fig.4)and those in bulk precipitation(4.36 kg·ha-1).Despite the model assumptions and limitations, this method remains effective in distinguishing internal and external sources in net throughfall flux in subalpine forest ecosystems.

Dry deposition is one of the main ion sources in throughfall(Balestrini et al., 2007; Malek and Astelet, 2008). Our results showed significant seasonal variations in dry deposition between gap edge canopies and closed canopies. Balestrini et al. (2007) and Erisman and Draaijers(2003) documented that the distance from the sea, tree height, wind speed, canopy cover, and surface conditions could be the main drivers that influenced the amount of dry deposition.Among them,the leaf area index is positively linearly related to dry deposition fluxes (Zirlewagen et al.,2001;Erisman et al.,2003).Therefore,the relatively sparse canopy in the gap-edge canopy is less effective in capturing dry deposition,which leads to lower annual dry deposition fluxes of Mg2+,K+and Ca2+in the gap-edge canopy than those under the closed canopy (Fig. 5).Seasonally,the dry deposition fluxes of K+,Ca2+and Mg2+were higher during the rainy season than during the snowy season for both canopies.The close reason for this observation may be related to the higher humidity of canopy leaves in the rainy season,which can be convenient for capturing more particles from the atmosphere(Gautam et al.,2017;Jiang et al.,2021).Identical studies have also been conducted in various forest ecosystems and demonstrated the significance of ion dry deposition in the rainy season(Devlaeminck et al.,2005;Zhang et al.,2007;Sun et al.,2021). Meanwhile, when the temperature decreases in winter, the physiologically deciduous L. mastersiana-A. faxoniana (Fu et al., 2017)can reduce the area of the canopy surface and limit the intercept of dry deposition(Gautam et al.,2017).In addition,significantly enhanced dry deposition fluxes of K+,Ca2+,Na+and Mg2+in the closed canopy were detected compared to those in the gap edge canopy (2.18–8.00-fold higher during the rainy season and 2.05–3.00-fold higher during the snowy season). The seasonal changes in soil nutrient availability may also be a reason (Hofhansl et al., 2011).

Canopy exchange is a bidirectional ion exchange process between intercepted precipitation and canopies with foliar or epiphytes (Van Langenhove et al.,2020).Previous studies have demonstrated that K+is considered a highly mobile ion that easily leaches from plant tissues by precipitation (Parker, 1983; Fan and Wei, 2001). However, the annual canopy exchange fluxes of K+in the closed canopy displayed negative values,which suggests the net absorption of the canopy,although a small amount (0.01 kg·ha-1) of K+leaching was also observed under the gap-edge canopy. The likely explanation is the epiphytes in canopies,which may incorporate nutrients into their thallus from precipitation and absorb nutrients for their growth (Knops et al., 1996). Both absorption and leaching of Ca2+have been reported among different forests(Jordan et al., 1980; Scheer, 2011; Van Stan and Pypker, 2015; Mukesh et al.,2017; Tonello et al., 2021). Ca2+is closely associated with leaf transpiration and other plant water metabolisms (McLaughlin and Wimmer,1999). The gap edge canopy and closed canopy showed obvious Ca2+absorption in the present study,which implies that canopy absorption of Ca2+may be a nutrient conservation strategy for forests when nutrients are restricted (Jordan et al., 1980). In contrast, Mg2+was leached from both canopies in our study.Overall,the gap edge canopy causes a lower absorption or leaching rate of base cations than that in the closed canopy.Principally,the nutrient availability at the site can be an important factor that determines nutrient leaching or absorption from the leaves when precipitation passes through the canopy, which represents a self-regulating process(Jordan et al.,1980).

The seasonal changes in plant physiological characteristics can also regulate the canopy exchange processes of K+, Ca2+and Mg2+(Tob′on et al., 2004; Hofhansl et al., 2011). In the rainy season, both Ca2+and Mg2+were absorbed,and K+was leached in the gap edge canopy,while K+, Ca2+and Mg2+were taken up in the closed canopy. However, the gap-edge canopy and closed canopy took up K+and Ca2+but leached Mg2+in the snowy season(Fig.5).Moreover,the canopy exchange fluxes of base cations were most often lower in the gap-edge canopy than in the closed canopy, especially in the snowy season, which may be partly related to the low leaf physiological activity of trees in the canopy in winter.Thus,the canopy can minimize the leaching of leaves and absorb nutrients from precipitation(Jordan et al.,1980).

There are uncertainties in our estimation of the throughfall deposition of base cations in subalpine natural forests. On one hand, the assessment must focus on the chemical properties of leaves, such as seasonal changes in water content, nutrient status of leaves, leaf transpiration,and soil nutrient limitation(Levia,2011;Van Stan et al.,2020),which can be more helpful to reveal the internal mechanism of the exchange process(i.e., gap edge canopy and closed canopy).On the other hand, since clouds and fog frequently occur in high-altitude areas, our analysis did not account for the deposition of base cations caused by clouds and fog between the gap edge canopy and closed canopy due to technical constraints. Even so, this study emphasizes that the gap edge canopy is a vital landscape position of K+, Ca2+, Na+and Mg2+from throughfall deposition in the studied forest. Therefore, future research must consider the driving factors that affect the gap edge canopy, as discussed above,to better guide the sustainable development of forests.

5. Conclusions

In summary, the concentration and fluxes of base cations in throughfall at the gap edge canopy were less than those in the closed canopy, especially during winter. The dry deposition and canopy exchange in both canopies were estimated by a canopy budget model,which shows that the annual canopy exchange fluxes of K+and Mg2+were lower in the gap edge canopy than in the closed canopy.However,the dry deposition fluxes of base cations significantly decreased from the closed canopy to the gap-edge canopy regardless of seasonal changes.The results suggest that the presence of a forest edge canopy leads to less base cation input from the throughfall deposition, and the fluxes of K+,Ca2+, Na+and Mg2+in the forest may be overestimated. However, the gap-edge canopy effect on the base cation input has not been well documented or underrepresented in forest hydrological research.

Funding

This study was financially supported by the National Natural Science Foundation of China(Nos.31922052,32022056 and 32171641).

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

Siyi Tan and Qing Dong designed this research,conducted the fieldwork and wrote the main manuscript;Shu Liao assisted in data analysis;Fuzhong Wu,Xiangyin Ni and Kai Yue designed this research.