Impact of Spatial and Environmental Variables on Aquatic Macroinvertebrates in Hilly Ponds


 Ponds are “islands” of independent habitats and research hotspots of regional biodiversity. This paper got some basic information about habitat conditions and aquatic macroinvertebrates in hilly ponds. Then we tried to provide a theoretical basis for the regional biodiversity conservation and sustainable development of the pond networks among basins. In this paper, the environmental and spatial variables concerned about aquatic macroinvertebrates in the ponds with different small basins of Liangping District, which is located in the east of Sichuan Basin, China, were investigated. The results showed that drainage basin effect did not occur among basins. The aquatic macroinvertebrates in the hilly ponds were mainly affected by altitude, nutrients, pH and other environmental factors. Within the basin scale, pond isolation was also one of the important factors affecting community structure. Considering the spatial pattern of the macroinvertebrate assemblages in the hilly ponds, the protection and management of those ponds need to focus on pond network other than a single pond, improve the habitat quality of the ponds, and ensure the cost-effectiveness and ecological benefits.


Introduction
As the complex terrain, changeable climate and extremely uneven rainfall in hilly areas of China, agriculture was limited and often depended on the utilization of water infrastructures in these areas. Due to the limitation of terrain, it is di cult to build large-scale water infrastructures in hilly areas, while it is more suitable to build small and scattered water infrastructures. Thus a large number of ponds as water infrastructures are constructed as important facilities for agricultural production. In addition to irrigation, ponds are also used for aquaculture and recreation etc.. In recent years, more attention has been paid to However, the understanding of ponds in hilly areas is obviously insu cient. It is still necessary to understand the characteristics of the assemblages and its impact factors.
Freshwater organisms are disappearing sharply around the world, and many aquatic ecosystems and its assemblages in the region are becoming more and more similar or homogeneous (Petsch, 2016).
However, spatial and temporal variations on biodiversity of aquatic organisms, especially for aquatic macroinvertebrates, had been poorly studied within basin (McLean, Mushet, Sweetman, Anteau, & Wiltermuth, 2020). Ponds are scattered and relatively independent freshwater ecosystems that support many important species of macroinvertebrates. The important driving forces for the composition of these assemblages include environmental and spatial factors (Free et al., 2009) A key challenge for biogeography and conservation biology is to understand spatial patterns and environmental drivers of freshwater biodiversity and community structure (Soininen, 2016). The ignorance about spatial process and environmental lters of aquatic macroinvertebrate assemblages in hilly ponds with lower altitude prevents regional ecological protection. A large number of natural ponds in hilly area have been declining due to over exploitation of human. Whether arti cial ponds can offset the ecological losses caused by the disappearance of natural ponds need to be studied as soon as possible.
Those studies determine the biodiversity and conservation values of these neglected and increasingly threatened habitats. To explore the composition and structure of aquatic macroinvertebrate assemblages in the ponds of hilly area and its relationship with environmental and spatial lters, it is hypothesized that: (1) Drainage basin effect of macroinvertebrates from the arti cial ponds in hilly areas exists among different basins; (2) The aquatic macroinvertebrate assemblages in the arti cial ponds are mainly driven by altitude, nutrients and pH; (3) Spatial processes can affect the diffusion and distribution of aquatic macroinvertebrate assemblages in hilly area. The results can help us understand the community characteristics and its impact factors of aquatic macroinvertebrates in the hilly ponds, provid basic data for regional ecological protection in hilly areas. Additionally, we can offer reasonable insights of basin management for pond managers.

Study area
Liangping district is located in the parallel ridge valley area in the east of Sichuan Basin. It belongs to subtropical monsoon climate with high humidity. The pH value of soil is 5.5-7.5. The main characteristics of rivers in Liangping are short length, small catchment and unstable runoff. Precipitation is the main resource of the surface water in Liangping.

Basin division and sampling sites
Basin division was based on DEM data with 1 m spatial resolution in Liangping district by ArcGIS 10.4.1 ( Fig. 1 (d)). According to the basin division map, three different basins were selected, and then 40 ponds were randomly selected as sampling sites in each basin to get environmental and spatial information (Fig. 1). A total of 120 ponds were collected with both water samples and aquatic macroinvertebrate samples. The sampling time was June 24-29, 2020 (late spring). The sampling ponds were mainly constructed in the process of agricultural production and living, which were arti cial or semi-arti cial ponds, and usually with permanent inundation. The main functions of those ponds were aquaculture, irrigation, bait sh culture and landscape, etc..

Environmental and spatial data
Environmental variables: altitude and area data were extracted in ArcGIS based on DEM data and remote sensing image map with resolution of 1m in Liangping district. The transparency of sampling sites was measured by Secchi disk. During the campaign, the coverage of aquatic macrophytes was estimated, and the information of the main functions, aquaculture situation, water depth were investigated by the manager of each pond. A 1L water sampler was used to collect water samples at different sites of each pond for 5-8 times, and then stored in a 10L sampling container. A 500 ml sampling bottle of water was sampled in the 10 L sampling container and stored in a 4 ℃ incubator. Then dissolved oxygen (DO), water temperature (T), pH and conductivity were measured in the 10 L sampling container by HACH

Macroinvertebrate eld surveys
The sampling of aquatic macroinvertebrates was carried out by a pond net (diameter, 30cm; 0.25cm, mesh size). The samples were collected in different microhabitats of each pond for 3 min (Briers & Biggs, 2005). Then the samples were stored in 80% alcohol and brought back to the laboratory. The samples were identi ed under microscope, and identi ed to the lowest level by using relevant taxonomic keys (Keast, 1994), or consulting relevant experts.

Statistical methods
One-way ANOVA was used to identify the signi cance of environmental and spatial variables among different basins. One-way ANOVA was both performed by IBM SPSS Statistics 25. The community structure of aquatic macroinvertebrates was represented by taxon richness, the Shannon-Wiener diversity index and relative abundance. Principal component analysis (PCA) was used to compare the community structure similarity of aquatic macroinvertebrates in different ponds. Principal component analysis was performed in Past3. Canonical correlation analysis (CCA) was used to analyze the correlation between variables and aquatic macroinvertebrates. Before canonical correlation analysis, the data of aquatic macroinvertebrates were analyzed by detrended correspondence analysis (DCA) to con rm a unimodal rather than linear distribution. Then environmental variables that had signi cant correlation with macroinvertebrate assemblages were selected by a Monte Carlo permutation test. Canonical correlation analyses were carried out by the package CANOCO 5.

Environmental and spatial variables
The results of environmental and spatial variables of the ponds in Liangping district are shown in Table 1. Total phosphorus had no signi cant differences among the three basins, and the concentration of TP ranged from 0.09 to 0.55 mg/L. The concentration of TN in basin 2 was signi cantly lower than that of in basin 1 and basin 3, and the concentration of TN in three basins was 0.68-7.74 mg/L. Total nitrogen was mainly composed of ammonium and nitrate. The concentration of ammonium in basin 2 was also signi cantly lower than that of in basin 1 and basin 3. The concentration of ammonium in basin 2 ranged from 0.15 to 1.88 mg/L. There were signi cant differences in water temperature, DO and transparency among the three basins. The highest value of water temperature, DO and transparency were in basin 1, and the lowest values of those three variables were in basin 3. The water temperature of basin 1, 2 and 3 were 27-36.

Spatial pattern of macroivertebrates
A total of 46 aquatic macroinvertebrates were collected, that belonged to 25 families and 41 genera and 46 species (Table 2). Forty-one species were collected in basin 2, while 32 and 31 species were collected in basin 1 and basin 3, respectively. Nine taxa of Gastropoda, one taxon of Crustacea and thirty-six taxa of Insecta were collected in this experiment. Thirteen taxa of Insecta belonged to Chironomidae. Those taxa of Parafossarulus sp., Dytiscus marginalis, Gyrinidae sp., Tanytarsus sp., Chironomus kiiensis, Micropsecra sp. and Eukiefferiella brehmi were only collected in basin 2.  Table 3, the mean taxon richness and Shannon-Wiener index of basin 3 in Liangping district were signi cantly lower than those in basin 1 and basin 2. The taxon richness of aquatic macroinvertebrates in ponds of basin 1 ranged from 3 to 15. The taxon richness of aquatic macroinvertebrates in ponds of basin 2 ranged from 3 to 19. The taxon richness of aquatic macroinvertebrates in ponds of basin 3 ranged from 3 to 14.

Relationship between variables and aquatic macroinvertebrates
The longest length of the DCA axis for aquatic macroinvertebrate data in basin 1, 2 and 3 were 3.54, 3.53 and 3.73, respectively. The results indicated that CCA can be used to characterize the relationship between macroinvertebrate assemblages and variables with signi cant relationship. The rst axis of DCA explained 16.61%, 11.61% and 17.17% of aquatic macroinvertebrate pro les in basin 1, 2 and 3, respectively. The other three axes of DCA explained 20.18%, 21.09% and 22.63% of the variables in basin 1, 2 and 3, respectively. There were 6, 3 and 2 variables in basin 1, 2 and 3 respectively, which had signi cant correlation with aquatic macroinvertebrate assemblages (p ≤ 0.05). According to CCA plot of basin 1, the eigenvalues of the rst axis and the second axis were 0.1628 and 0.1257, respectively. Environmental and spatial variables accounted for 23.7% of the variation information of species composition. For CCA of basin 2, the eigenvalues of the rst axis and the second axis were 0.1679 and 0.1256, respectively. Environmental and spatial variables accounted for 13.7% of the variation information of species composition. According to CCA of basin 3, the eigenvalues of the rst axis and the second axis were 0.1689 and 0.0405 respectively. Environmental and spatial variables accounted for 8.9% of the variation information of species composition.
In CCA plot of basin 1 (Fig. 5 (a) In basin 1, the species composition was also affected by DO, and the species had positive correlation with DO included Stratiomys sp., Dicrotendipes sp., etc.. In the CCA plot of basin 2 ( Fig. 5 (b)), altitude, area and pond isolation had signi cant correlation with species composition. The gradient of altitude among sampling sites in basin 2 was larger than that of in basin 1 and 3. The results showed that there was a positive correlation between altitude and Dicrotendipes tritomus, Culicidae sp., Baetis sp., Pseudothemis sp.. Meanwhile those species had negative relationship with pond isolation. Chironomus sp., Physa sp., Galba pervia showed positively correlation with area. In CCA plot of basin 3 (Fig. 5 (c) . The taxa collected in this experiment were less, which might be due to the high concentration of nutrients in the ponds. The diversity of aquatic macroinvertebrates was usually negatively correlated with the concentration of nutrients. We found that the concentration of TN was higher than 2 mg/L, and the concentration of TP was higher than 0.2 mg/L in most ponds, which meant on the state of eutrophication. Figure 5(a) also showed that a large number of invertebrates were inhibited by nitrogen. In this experiment, the major function of most ponds was aquaculture. The management behavior of managers, such as feeding and disinfection, would also have a strong interference on the aquatic food chain, thus reducing its biodiversity. Therefore, the results of this experiment identi ed a subset of aquatic macroinvertebrate assemblages in hilly region, and also showed that the value of the regional biodiversity was damaged. It had great potential to improve the habitat quality and ecosystem services of the ponds.

Spatial pattern
With the increase of physical distance, the limitation of taxa diffusion would increase, which might form spatial correlates (Legendre and Legendre, 1998). Therefore, we usually assumed that habitats with closer distance had more similar communities than habitats that were far apart, and the intensity of their impact depended on the characteristics of individual diffusion ability (Razeng et al., 2016). For the study of freshwater organisms, the study area might be a single basin or basins. It was generally believed that biological diffusion was more likely to occur within a basin than among basins (Jani Heino et al., 2015). Therefore, the spatial correlates of species are more likely to occur among basins. Aquatic assemblages can be affected by both spatial and environmental variables. However, it is not clear at what spatial scale, spatial correlates is more likely to occur. It is generally believed that larger spatial scales are more prone to spatial correlates. . This supports the hypothesis that a common typology can be applied to a group of aquatic ecosystem with irregular distribution in a large-scale region. The investigation area of this experiment was 1892.13 km 2 , and the spatial correlates of taxa were not obvious. There are two possible reasons for this phenomenon. One is that compared with other spatial correlates studies, the scale of the study area is smaller, the geographical barrier of species among basins is smaller, and it is easy for macroinvertebrates to diffusion. Although geographical factors were the main factors controlling the composition structure of macroinvertebrates in wetlands, but the differences among basins in smallscale region were negligible (Batzer & Ruhi, 2013). Second, the altitude of the study area is low.

Impact of environmental and spatial variables on the assembles
In large-scale research, spatial correlation is more likely to occur. However, in small-scale and scattered wetlands, the spatial variation of aquatic macroinvertebrate community was mostly explained by local Wilson 1967), it could be concluded that high quality and large ponds had higher biodiversity, including biodiversity of invertebrates. According to the habitat diversity hypothesis, the impact of area on species richness is mainly realized through biodiversity. The larger the area, the greater the species diversity (Ricklefs & Lovette, 1999). The habitats with large area, permanent inundation and alkalinity tended to have high biodiversity (Hoverman et al., 2011). However, some studies had shown that a group of small ponds was a key contributor to the biodiversity of regional invertebrates (Boix et  were collected in acidic ponds (pH 6 ≤ pH < 6.6), and 8 gastropods were collected in acidic pond system (pH 4.4-6) (Spyra, 2017). When the pH value of water was lower than 5.2, gastropods might not survive (Singh & Agrawal, 2008). There was no reference value for the pH tolerance range of gastropods, and it was considered that 5.5-9.5 was the adaptive range of gastropods (Spyra, 2017).
Three aspects show the effect of pH on invertebrates in ponds as follows. First, a large number of studies had shown that the higher pH value was, the greater taxa richness of invertebrates in ponds was (J. values were associated with higher residues and humus of vegetation (Vuorenmaa, Forsius, & Mannio, 2006). The accumulation of detritus also increased the diversity of habitat structure, thus increasing species richness (Schmidl, J, 2003). Thirdly, some studies found that pH value had no signi cant effect on density of invertebrates (Simpson, Bode, & Colquhoun, 1985;Winterbourn & Collier, 1987). Therefore, under certain conditions, pH will have an impact on aquatic invertebrates, and the impact on different species is different. Some invertebrates such as Odonata and aquatic beetles may prefer alkaline water, while some invertebrates such as aquatic beetles and Oligochaeta may prefer acidic water. In this paper, there was a signi cant positive correlation between pH and Spercheus emarginatus which belonged to aquatic beetles, and a signi cant negative correlation between pH and Diplonychus esakii which belonged to aquatic beetles.
(  (Bronmark, 1985) and the migration pattern of predatory aquatic insects (Wilcox, 2001). Briers and Biggs believed that the isolation degree of ponds had a signi cant impact on invertebrate community, and the invertebrate community structure among adjacent sampling sites was more similar than that among distant ponds (Briers & Biggs, 2005). Signi cant correlates between isolation degree and species assemblageswere found in both basin 1 and basin 2 which was in line with previous studies.

Implication for biodiversity conservation of ponds
Different aquatic ecosystems support different community structures of aquatic macroinvertebrates. The physical and chemical characteristics of water, such as pH, DO and transparency, are important factors affecting the community structure of aquatic macroinvertebrates. At the same time, due to the regularity of regional water quality, regional aquatic macroinvertebrates usually show relevant regularity. So we can understand the characteristics of regional biodiversity, the scale and species to be protected. To a certain extent, arti cial ponds of hilly area in China maintain the scale of local aquatic species, expands the habitat area, and improves the functional connectivity of most ponds. As a "stepping stone" of habitat, it provides convenience for the migration and diffusion of aquatic organisms. However, hilly ponds are also faced with over utilization and serious agricultural non-point source pollution. In addition to the establishment of statutory reserves, we need to shift the unreasonable utilization of ponds to the sustainable development of biodiversity resources desperately. The cost-effectiveness of protection management must be considered, and protection decisions also be constrained by multiple con icts of interest. Therefore, we need to conserve the habitat and biodiversity of ponds from the aspects of landscape scale, habitat conditions and industrial structure etc.. temperature and climate (such as altitude), habitat structure (such as area). At the same time, we should be alert to the loss of biodiversity caused by acidi cation of human activities. In China, a large number of arti cial ponds are permanent, and they may become temporary wetlands for no arti cial management. As a part of the protection and management strategy, it was encouraged to maintain the pond network with different length of hydrological cycle and environmental characteristics (Hill et al., 2017). In terms of industrial structure, we can introduce more sustainable utilization modes into the production and utilization of ponds, increase the economic output of ponds, and realize biodiversity conservation with low-cost.

Conclusion
Hilly ponds support special biological assemblages and contribute to regional biodiversity. However, in this paper, the habitat quality of the ponds was damaged, resulting in the low taxa richness of the aquatic macroinvertebrate assembles. It found that there was little difference of the macroinvertebrate assemblages among basins. These hilly ponds were mainly affected by local environmental factors and spatial processes, including altitude gradient, nutrients, area, temperature, pH and pond isolation etc.. On the premise of understanding the local ponds habitat conditions and aquatic assemblages status, we put forward the strategy of biodiversity conservation of regional ponds. We should focus on the pond network, control the cost-effectiveness and increase the ecological bene ts by developing sustainable wetland economy.  The faunal composition of aquatic macroinvertebrates in each basin. Figure 3