Effects of sampling on the elevational distribution of nematode-trapping fungi CURRENT STATUS: UNDER REVISION

Background Modelling species richness across an elevation gradient has long attracted attention, and at same time places some significant obstacles to research. Many interpretations of patterns and corresponding mechanisms for species distributions are made without consideration of multiple confounding factors. What are factors that affect species richness with elevation? The answer may contribute to better understanding of the elevational distribution patterns and mechanisms.In this study, we performed the research on species richness of nematode-trapping fungi (NTF) across an elevation gradient in Yunnan, China. Results The results showed that sampling patterns, sampling altitude range, and human disturbance in sampling site could affect the resulting patterns of species richness significantly. Conclusion The results suggested that future studies on the elevational gradients of species richness should address these factors, and try to adopt the high-sampling patterns to reduce the observation bias.


Background
In mountainous environments, steep environmental gradients can occur with slight variances in elevation. Ecological patterns observed across elevation gradients has contributed to the underpinning of ecological niche theory and the current knowledge of global life zones, species composition and biogeography [1]. Elevational gradients remain a mainstay for contemporary research on many topics in ecology and evolution and are critical to our current understanding of large-scale trends in biodiversity, global change and conservation.
The mid-elevation peak is the most common pattern observed [5,18].
Studies have offered numerous factors that regulate species richness with changing elevation, such as size and scale of the studied site [2,19]. However, in recent years, researchers have identified other important factors such as sampling pattern, sampling range and human disturbance. Nogués et al (2008) showed that species richness patterns were influenced by the altitude of areas studied and levels of human disturbance [20]; Grytnes (2008) showed that sampling intensity also affected the results of elevational gradient study, using a rarefaction curve to circumvent the resulting errors [21]. However, at present, neither the effects of sampling patterns, sampling range nor the influence of human disturbance have been sufficiently identified.
The Three Parallel Rivers Region is a world hotspot for biodiversity. The region has dramatic extremes in elevation (up to 6000m) from the Nujiang, Lancang and Jinsha Rivers to the upper limits of the adjacent mountain ranges making it an outstanding area for testing theories of elevational gradients. In this study, we examined the vertical distribution pattern of nematode-trapping fungi (NTF) richness at two sites (Gaoligong Mountain and Cangshan Mountain) and identify the regulating factors to richness. We 4 believed comparing these sites would demonstrate that the absolute elevation range and sampling layout would influence the richness pattern observed.

Studied subjects
NTF are a type of predatory Eukaryotic microorganisms that capture nematodes with specialized vegetative mycelia that function as trapping structures. These trapping structures include adhesive network (Arthrobotrys), adhesive knobs and branches (Dactylellina), and constricting rings (Drechslerella). NTF are widespread, occupying a range of aquatic and terrestrial habitats, and are an important natural biocontrol factors [23]. The limited species number and relative ease to identification make NTF ideal to ecological research.

Studied areas
This field study was conducted at two sites (Gaoligongshan and Cangshan) in the Three

Sample collection
Soil samples were collected using two modes ("distance interval method", "elevation

Preparation of culture medium
Corn meal agar medium was used to isolate and purify the NTF; Potato dextrose agar medium was used to extract strains for eutrophic culture [22].

The isolation, purification and identification of NTF
Soil sprinkling technique were used to isolate and purify the NTF from the soil samples, and identify the NTF species in terms of morphology and molecular biology [22,23].

Data treatment
Occurrence frequency (OF) = the quantity of soil sample for individual species/total soil samples x 100%.
The original data was sorted and analysed by EXCEL and SPSS programme. 6

The elevational distribution pattern of NTF in Cangshan Mountain
At Cangshan, there were 3 genera and 12 species and 57 strains of NTF isolated and identified from 300 soil samples, the OF decreased with increasing elevation, consistent with the decreasing pattern. The species number initially increased and subsequently 7 decreased, with the peak value biased towards low altitude, conforming to LPMP pattern ( Fig 5).

The elevational distribution of NTF after removal of the human interference range
The low-altitude areas of eastern slope and the western slope of Cangshan Mountain, to some degree, were disturbed by human activities, but the ecological function areas of Cangshan Natural Reserve were less disturbed. The sampling areas in this study, located at the altitude of 2100-2300 meters, were outside the function areas, thus it was more affected by human activities.
Excluding data outside the altitude range of 2100-2300 meters, rarafaction curve and

Discussion
The accuracy of results on studying the vertical distribution patterns of NTF may be affected by sampling patterns, scale of study and post-sampling treatment of data.
Geographical limitation, sampling uniformity, habitat fragmentation and human impacts can also influence distribution patterns [8,20,24]. Although previous studies have acknowledged these impacting factors, few have been validated by examples. This study on the elevational distribution patterns of NTF in the Three River Parallel Regions, have suggested that sampling pattern, sampling range, human disturbance and multidimensionality of biodiversity all affected the research outcome.

Sampling modes affecting the elevational distribution pattern of NTF
In this study, the different sampling methods have obtained different results on the elevational distribution patterns of NTF. In the distance sampling method, the OF of NTF conformed to midpeak pattern, and species number met LPMP pattern; however, the elevation sampling methods obtained a totally different result that the OF and species number of NTF both showed a decreasing pattern. It suggested that sampling methods would affect the observed pattern. Our results emphasize the importance of sampling method in developing species richness and OF models along environmental gradients.
Undoubtedly, in the process of studying the distribution patterns of species, increasing sampling efforts sees a consequent increase to a model's accuracy [25]. However, when the sampling sites are unevenly distributed, the results will likely be biased. The equidistantly sampling method along the sample line in Gaoligong Mountain resulted in the fact that the sampling points were not averagely distributed on the elevational gradient, but concentrated in the middle altitude range, which probably caused the overestimation on the species richness of NTF in this area. The correlation analysis on the number of species and sample points indicated that they were strongly correlated with each other, thus this sampling pattern in which the sample points were concentrated in the middle altitude areas showed a mid-elevation peak pattern.
When using the evenly sampling method along the elevational gradient in Gaoligong Mountain, the elevational distribution of NTF showed a decreasing pattern, this method would not be affected by the distribution of sampling sites and the scale of studied region, thus the decreasing pattern was probably closer to the true situation of the elevational distribution of NTF. We used the same sampling method in Cangshan Mountain, and got the same results that NTF OF showed a decreasing pattern. This demonstrated that our hypothesis was validated in different regions, and the elevation sampling method can 9 resolve the observation biases and present more accurate results.

Sampling range affecting the elevational distribution of NTF
When we truncated the sampling range to 2100-3500 meters in the Gaoligong Mountains where the samples were equidistantly collected along the sample lines (distance method), it was found that the original mid-elevation peak pattern (mid-peak pattern for OF, LPMP pattern for species number) changed to a decreasing pattern for the elevation sampling method. Previous studies on truncation and scale effects on the vertical distribution of species have been carried out. Truncating the low-altitude range of the studied region led to the changes of the vertical distribution of species from the midpeak pattern to a decreasing pattern [20]. When the elevational gradient was entirely surveyed, the pattern was hump-shaped, changing eventually to a decreasing pattern as the scale of extent diminished. Likewise, the OF at the same altitude range of Cangshan Mountain (2100-3500 meters) was also in line with the decreasing pattern. This result further supported that gradient truncation significantly affected the research on the elevational distribution patterns, this also confirmed our speculation that the sampling range affected the research and the diminishing of sampling range caused the changes of mid-peak pattern to the decreasing pattern.

Human disturbance affecting the study on the elevational distribution of NTF
Areas rich in biodiversity often overlap with areas of high human populations, and it's generally accepted that human disturbance can affect the distribution of biodiversity [26][27][28]. Surprisingly the influence from human disturbance on species richness models has been long ignored. In recent years, some researchers have given this some attention, pointing out that human populations are generally based around low elevations, and therefore human disturbance is not evenly distributed across most elevations [29][30].
The OF of NTF in Cangshan Mountain was in the decreasing pattern, while the species number showed the LPMP pattern. In eastern slope of Cangshan Mountain, below the altitude of 2200 meters is residential area, in west slope of Cangshan Mountain, below 2400 meters is farmland, the boundary of Cangshan Nature Reserve is in the range of 2100-2300 meters, thus the areas outside the boundary may be affected by varying degrees of human disturbance [31]. When omitting the data collected in the range of 2100-2300 meters of Cangshan Mountain, we found both the OF and species number of NTF in a decreasing pattern, which evidently proved that human disturbance affected the vertical distribution of NTF. The declined biodiversity in the low-altitude areas because of human disturbance led to the peak of biodiversity in the middle-altitude areas. This was probably the reason why the species number of NTF in Cangshan Mountain showed a LPMP pattern when including the human-disturbed areas.

Prospects and suggestions
When we used different indicators (OF and species number) to identify microbial biodiversity on the altitude gradient in this study, the same set of studies presented different results, it probably meant that our previous studies did not reflect the full picture of biodiversity. Johnathan et al (2018) pointed out that because of multidimensional and scale-dependent characteristics of biodiversity, it would be better to describe its change from multiple perspectives [32]. Multidimensionality made the study of biodiversity at different time and different space more challenging than other variables in ecology [33]. In some studies, elevational distribution patterns present by used different biodiversity indicators (species number, species density, evenness, biodiversity index) are different [17,[34][35], thus the distribution patterns shown by different biodiversity indicators seemed to be interlaced, and different indicators showed different dimensions and different levels of biodiversity, to describe the biodiversity more fully from multiple dimensions needs more research and exploration.
Despite of the viewpoints from Nogues et al (2008) that the removal of high altitude areas had little effect on the research of the vertical distribution of species [20], Three River Parallel Region in China is characterized by vast elevation span, vertical climate and vertical vegetation, the alpine mudstone beach and yearly snow-covering in the highaltitude areas may cause steep fall of biodiversity. In this study, the altitude range was not large enough to completely cover the whole range of "the Three Rivers Parallel Region" and to carry out the exploration of the impact on the vertical distribution in the high-altitude areas. Future research should also focus on the integrity of the altitude range and further extend the studied areas.
The results in this study suggested that future study should address the sample-laying patterns and ensure not only the sampling evenness at each altitude, but also the consistency of the altitude range between the studied areas. Whenever and wherever possible, rarefaction curve should be used to determine the rationality of the sampling, at the same time, attention should also be paid to the impact of human disturbance on the elevational distribution patterns of species. We also noted that the OF was frequently used in the microbiology study as an indicator of species density which was relatively less affected, but the large animals might be more affected if the number of species was used as an indicator, therefore, the species inconsistency which was in the studies on the vertical distribution may be related to sampling patterns. Of course, the environmental heterogeneity caused by altitude was more than the difference of temperature, precipitation and vertical area. In the future, we need to systematically carry out the research by integrating environment, biological groups, sampling patterns and data analysis to obtain the real situation of the vertical distribution of species, which is crucial to understand the forming mechanism, maintaining mechanism and large-scale distribution of microorganisms.

Conclusions
The resultsshowed that sampling patterns, sampling altitude range, and human disturbance in sampling sites do affected the understanding on NTF elevational distribution. Our study suggested that future studies should concerned more on sampling design, agreement on biodiversity elevational distribution can be achieved only after agreement on sampling is made. The elevational distribution pattern of NTF in the smaller altitude range of sampling.

List Of Abbreviations
21 Figure 5 The elevational distribution pattern of NTF in Cangshan Mountain.
22 Figure 6 Elevational distribution patterns of NTF in two regions of 2300-3500m and their correlation.